enlightenment/src/menu.c

2133 lines
42 KiB
C
Raw Normal View History

Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
#include "debug.h"
#include "menu.h"
#include "config.h"
#include "util.h"
#include "math.h"
2000-12-29 12:49:17 -08:00
2003-01-15 10:29:56 -08:00
static Evas_List * open_menus = NULL; /* List of all open menus */
static Evas_List * menus = NULL;
static E_Menu_Item *curr_selected_item = NULL; /* Currently selected item */
static Window menu_event_win = 0; /* Window which originated event */
static int screen_w, screen_h; /* Screen width and height */
static int mouse_x, mouse_y; /* Mouse coordinates */
static int keyboard_nav = 0; /* If non-zero, navigating with keyboard */
static void e_idle(void *data);
static void e_wheel(Ecore_Event * ev);
static void e_key_down(Ecore_Event * ev);
static void e_key_up(Ecore_Event * ev);
static void e_mouse_down(Ecore_Event * ev);
static void e_mouse_up(Ecore_Event * ev);
static void e_mouse_move(Ecore_Event * ev);
static void e_mouse_in(Ecore_Event * ev);
static void e_mouse_out(Ecore_Event * ev);
static void e_window_expose(Ecore_Event * ev);
static void e_menu_item_select(E_Menu_Item * mi);
static void e_menu_item_unselect(E_Menu_Item * mi);
static void
e_scroller_timer(int val, void *data)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
int ok = 0;
int resist = 5;
int scroll_speed = 12;
static double last_time = 0.0;
double t;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
2003-01-15 10:29:56 -08:00
resist = config_data->menu->resist;
scroll_speed = config_data->menu->speed;
t = ecore_get_time();
if (val != 0)
scroll_speed = (int)(((t - last_time) / 0.02) * (double)scroll_speed);
last_time = t;
ok = 0;
if (mouse_x >= (screen_w - resist))
{
int scroll = 0;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if ((m->current.x + m->current.w) > screen_w)
scroll = m->current.x + m->current.w - screen_w;
}
if (scroll)
{
if (scroll > scroll_speed)
scroll = scroll_speed;
e_menu_scroll_all_by(-scroll, 0);
ok = 1;
}
}
else if (mouse_x < resist)
{
int scroll = 0;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (m->current.x < 0)
scroll = -m->current.x;
}
if (scroll)
{
if (scroll > scroll_speed)
scroll = scroll_speed;
e_menu_scroll_all_by(scroll, 0);
ok = 1;
}
}
if (mouse_y >= (screen_h - resist))
{
int scroll = 0;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if ((m->current.y + m->current.h) > screen_h)
scroll = m->current.y + m->current.h - screen_h;
}
if (scroll)
{
if (scroll > scroll_speed)
scroll = scroll_speed;
e_menu_scroll_all_by(0, -scroll);
ok = 1;
}
}
else if (mouse_y < resist)
{
int scroll = 0;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (m->current.y < 0)
scroll = -m->current.y;
}
if (scroll)
{
if (scroll > scroll_speed)
scroll = scroll_speed;
e_menu_scroll_all_by(0, scroll);
ok = 1;
}
}
if ((ok) && (open_menus))
ecore_add_event_timer("menu_scroller", 0.02, e_scroller_timer, val + 1,
NULL);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
UN(data);
}
2000-12-29 12:49:17 -08:00
static void
e_idle(void *data)
2000-12-29 12:49:17 -08:00
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
2000-12-29 12:49:17 -08:00
for (l = menus; l; l = l->next)
{
E_Menu *m;
2000-12-29 12:49:17 -08:00
m = l->data;
e_menu_update_base(m);
}
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
e_menu_update_shows(m);
}
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
e_menu_update_hides(m);
}
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
e_menu_update_finish(m);
}
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (m->first_expose)
evas_render(m->evas);
2000-12-29 12:49:17 -08:00
}
again:
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (m->delete_me)
{
e_object_unref(E_OBJECT(m));
goto again;
}
}
e_db_runtime_flush();
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
UN(data);
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_wheel - Handle mouse wheel events
*
* @ev: Pointer to event.
*/
static void
e_wheel(Ecore_Event * ev)
{
Ecore_Event_Wheel *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
if (e->win == menu_event_win)
{
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_key_down - Handle key down events
*
* @ev: Pointer to event.
*/
2000-12-29 12:49:17 -08:00
static void
e_key_down(Ecore_Event * ev)
2000-12-29 12:49:17 -08:00
{
Ecore_Event_Key_Down *e;
int ok;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
2000-12-29 12:49:17 -08:00
e = ev->event;
ok = 0;
if (e->win == menu_event_win)
ok = 1;
else
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if ((e->win == m->win.main) || (e->win == m->win.evas))
{
ok = 1;
break;
}
}
}
if (ok)
2000-12-29 12:49:17 -08:00
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
E_Menu *m = NULL;
E_Menu_Item *mi = NULL;
for (l = open_menus; l; l = l->next)
{
m = l->data;
if (m->selected)
{
mi = m->selected;
break;
}
}
2000-12-29 12:49:17 -08:00
if (!strcmp(e->key, "Up"))
{
keyboard_nav = 1;
e_menu_select(0, -1);
2000-12-29 12:49:17 -08:00
}
else if (!strcmp(e->key, "Down"))
{
keyboard_nav = 1;
e_menu_select(0, 1);
2000-12-29 12:49:17 -08:00
}
else if (!strcmp(e->key, "Left"))
{
keyboard_nav = 1;
e_menu_select(-1, 0);
2000-12-29 12:49:17 -08:00
}
else if (!strcmp(e->key, "Right"))
{
keyboard_nav = 1;
e_menu_select(1, 0);
2000-12-29 12:49:17 -08:00
}
else if (!strcmp(e->key, "Escape"))
{
keyboard_nav = 1;
for (l = menus; l; l = l->next)
{
m = l->data;
if (m->current.visible)
e_menu_hide(m);
}
}
else if (!strcmp(e->key, "Return"))
{
keyboard_nav = 1;
if (mi)
{
e_menu_callback_item(m, mi);
mi->selected = 0;
mi->menu->selected = NULL;
}
for (l = menus; l; l = l->next)
{
m = l->data;
if (m->current.visible)
e_menu_hide(m);
}
2000-12-29 12:49:17 -08:00
}
else
{
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_key_up - Handle key up events
*
* @ev: Pointer to event.
*/
2000-12-29 12:49:17 -08:00
static void
e_key_up(Ecore_Event * ev)
2000-12-29 12:49:17 -08:00
{
Ecore_Event_Key_Up *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
2000-12-29 12:49:17 -08:00
e = ev->event;
{
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_mouse_down - Handle mouse down events
*
* @ev: Pointer to event.
*/
static void
e_mouse_down(Ecore_Event * ev)
{
Ecore_Event_Mouse_Down *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
{
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_mouse_up - Handle mouse up events
*
* @ev: Pointer to event.
*/
static void
e_mouse_up(Ecore_Event * ev)
{
Ecore_Event_Mouse_Up *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
keyboard_nav = 0;
if (e->win == menu_event_win)
{
if (open_menus)
{
E_Menu *m;
m = open_menus->data;
if ((e->time - m->time) > 200)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
for (l = open_menus; l; l = l->next)
{
m = l->data;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/* Ensure that the item is actually selected and
* that the mouse pointer really is over it: */
if (m->selected)
{
/* Get the dimensions of the selection for use in
* the test */
double s_x, s_y, s_w, s_h;
2003-01-15 10:29:56 -08:00
evas_object_geometry_get(m->selected->obj_entry,
&s_x, &s_y, &s_w, &s_h);
if (INTERSECTS(m->current.x + rint(s_x),
m->current.y + rint(s_y),
rint(s_w),
rint(s_h),
mouse_x, mouse_y, 0,
0))
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
{
e_menu_callback_item(m, m->selected);
}
e_menu_item_unselect(m->selected);
break;
}
}
for (l = menus; l; l = l->next)
{
m = l->data;
if (m->current.visible)
e_menu_hide(m);
}
}
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_mouse_move - Handle mouse move events
*
* @ev: Pointer to event.
*/
static void
e_mouse_move(Ecore_Event * ev)
2000-12-29 12:49:17 -08:00
{
Ecore_Event_Mouse_Move *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
keyboard_nav = 0;
if (e->win == menu_event_win)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
mouse_x = e->rx;
mouse_y = e->ry;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
2003-01-15 10:29:56 -08:00
/* checkme
evas_event_move(m->evas,
e->rx - m->current.x, e->ry - m->current.y);
*/
evas_event_feed_mouse_move(m->evas,
e->rx - m->current.x, e->ry - m->current.y);
}
}
else
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
mouse_x = e->rx;
mouse_y = e->ry;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
2003-01-15 10:29:56 -08:00
/* checkme
evas_event_move(m->evas,
e->rx - m->current.x, e->ry - m->current.y);
2003-01-15 10:29:56 -08:00
*/
evas_event_feed_mouse_move(m->evas,
e->rx - m->current.x, e->ry - m->current.y);
}
}
e_scroller_timer(0, NULL);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_mouse_in - Handle mouse enter events
*
* @ev: Pointer to event.
*/
static void
e_mouse_in(Ecore_Event * ev)
{
Ecore_Event_Window_Enter *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
keyboard_nav = 0;
if (e->win == menu_event_win)
{
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_mouse_out - Handle mouse leave events
*
* @ev: Pointer to event.
*/
static void
e_mouse_out(Ecore_Event * ev)
{
Ecore_Event_Window_Leave *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
keyboard_nav = 0;
if (e->win == menu_event_win)
{
}
else
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if ((e->win == m->win.main) || (e->win == m->win.evas))
{
2003-01-15 10:29:56 -08:00
evas_event_feed_mouse_move(m->evas, -99999999, -99999999);
}
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_window_expose - Handle window expose events
*
* @ev: Pointer to event.
*/
static void
e_window_expose(Ecore_Event * ev)
{
Ecore_Event_Window_Expose *e;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e = ev->event;
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (e->win == m->win.evas)
{
m->first_expose = 1;
2003-01-15 10:29:56 -08:00
/* checkme
evas_update_rect(m->evas, e->x, e->y, e->w, e->h);
2003-01-15 10:29:56 -08:00
*/
evas_damage_rectangle_add(m->evas, e->x, e->y, e->w, e->h);
break;
}
2003-01-15 10:29:56 -08:00
e_menu_update_shape(m);
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
/**
* e_menu_item_unselect - Unselect a menu item.
*
* @mi: Pointer to the menu item to be unselected.
*/
static void
e_menu_item_unselect(E_Menu_Item * mi)
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
{
D_ENTER;
if ((mi) && (mi->menu->selected == mi))
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
{
mi->menu->selected = curr_selected_item = NULL;
mi->selected = 0;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
mi->menu->redo_sel = 1;
mi->menu->changed = 1;
}
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_menu_item_select - Select a menu item.
* NOTE - Assumes only one item can be selected at once,
* and unselects any previously selected menu item.
*
* @mi: Pointer to the menu item to be selected.
*/
static void
e_menu_item_select(E_Menu_Item * mi)
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
{
D_ENTER;
// e_menu_item_unselect(curr_selected_item);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
if (mi)
{
mi->menu->selected = mi;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
mi->selected = 1;
mi->menu->redo_sel = 1;
mi->menu->changed = 1;
curr_selected_item = mi;
}
D_RETURN;
}
/**
* e_menu_item_in_cb - Callback for when mouse enters a specific menu item.
* Attached by e_item_realize(). Selects menu item.
*
* @_data: Pointer to actual menu item structure.
* @_e: Evas
* @_o: Evas object
* @_b: ?????
* @x: ?????
* @y: ?????
*/
static void
2003-01-15 10:29:56 -08:00
e_menu_item_in_cb(void *_data, Evas * _e, Evas_Object * _o, void *ev_info)
{
E_Menu_Item *mi;
2003-01-15 10:29:56 -08:00
Evas_Event_Mouse_In *event_info = ev_info;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi = _data;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
e_menu_item_select(mi);
e_menu_hide_submenus(mi->menu);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
if (mi->submenu && mi->submenu->entries)
{
e_menu_move_to(mi->submenu,
mi->menu->current.x + mi->menu->current.w,
mi->menu->current.y + mi->y - mi->menu->border.t);
e_menu_show(mi->submenu);
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
UN(_e);
UN(_o);
2003-01-15 10:29:56 -08:00
UN(event_info);
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_menu_item_out_cb - Callback for when mouse leaves a specific menu item.
* Attached by e_item_realize(). Unselects menu item.
*
* @_data: Pointer to actual menu item structure.
* @_e: Evas
* @_o: Evas object
* @_b: ?????
* @x: ?????
* @y: ?????
*/
static void
2003-01-15 10:29:56 -08:00
e_menu_item_out_cb(void *_data, Evas * _e, Evas_Object * _o, void *ev_info)
{
E_Menu_Item *mi;
2003-01-15 10:29:56 -08:00
Evas_Event_Mouse_Out *event_info = ev_info;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi = _data;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
e_menu_item_unselect(mi);
D_RETURN;
UN(_e);
UN(_o);
2003-01-15 10:29:56 -08:00
UN(event_info);
}
void
e_menu_callback_item(E_Menu * m, E_Menu_Item * mi)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (mi->func_select)
mi->func_select(m, mi, mi->func_select_data);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_hide_callback(E_Menu * m, void (*func) (E_Menu * m, void *data),
void *data)
{
m->func_hide = func;
m->func_hide_data = data;
}
void
e_menu_item_set_callback(E_Menu_Item * mi,
void (*func) (E_Menu * m, E_Menu_Item * mi,
void *data), void *data)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi->func_select = func;
mi->func_select_data = data;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
/**
* e_menu_hide_submenus - Hide all menus except @menus_after.
* Assumes all menus after @menus_after in the list open_menus
* are submenus of @menus_after.
*
* @menus_after: All menus after this are hidden.
*/
void
e_menu_hide_submenus(E_Menu * menus_after)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
/* Loop thru all open menus: */
for (l = open_menus; l; l = l->next)
{
/* Found submenu, so now hide all remaining menus: */
if (l->data == menus_after)
{
l = l->next;
for (; l; l = l->next)
{
E_Menu *m;
m = l->data;
e_menu_hide(m);
}
break;
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
/**
* e_menu_select - Attempt to select the menu entry @dx entries across,
* and @dy entries down.
*
* @dx: Horizontal offset of new menu entry.
* @dy: Vertical offset of new menu entry.
*/
void
e_menu_select(int dx, int dy)
{
2003-01-15 10:29:56 -08:00
Evas_List *l, *ll;
int done = 0;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
/* Loop through all open menus, tile done or reached end */
for (l = open_menus; (l) && (!done); l = l->next)
{
E_Menu *m;
m = l->data;
/* If this is the selected menu: */
if (m->selected)
{
/* Go through the menu entries: */
for (ll = m->entries; (ll) && (!done); ll = ll->next)
{
E_Menu_Item *mi;
mi = ll->data;
/* Found the currently selected entry: */
if (mi->selected)
{
/* Vertical movement, up and down menu: */
if (dy != 0)
{
int ok = 0;
/* Only go up or down if entry exists to do so, */
/* and skip over separators: */
if (dy < 0)
{
for (; ll->prev; ll = ll->prev)
{
mi = ll->prev->data;
if (!mi->separator)
{
ok = 1;
break;
}
}
}
else if (dy > 0)
{
for (; ll->next; ll = ll->next)
{
mi = ll->next->data;
if (!mi->separator)
{
ok = 1;
break;
}
}
}
if (ok)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/* Unselect old and select new item: */
e_menu_item_unselect(m->selected);
e_menu_item_select(mi);
e_menu_hide_submenus(mi->menu);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/* If submenu, display it: */
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
if (mi->submenu && mi->submenu->entries)
{
e_menu_move_to(mi->submenu,
mi->menu->current.x +
mi->menu->current.w,
mi->menu->current.y +
mi->y -
mi->menu->border.t);
e_menu_show(mi->submenu);
}
}
}
done = 1;
}
}
/* Horizontal movement, into and out of submenus: */
if (dx != 0)
{
int ok = 0;
/* Only carry on if appropriate submenus exist: */
if ((dx < 0) && (l->prev))
ok = 1;
else if ((dx > 0) && (l->next))
ok = 1;
if (ok)
{
E_Menu_Item *mi = NULL;
E_Menu *mm;
/* Moving out of a submenu: */
if (dx < 0)
{
2003-01-15 10:29:56 -08:00
Evas_List * ll;
mm = l->prev->data;
for (ll = mm->entries; (ll) && (!mi); ll = ll->next)
{
E_Menu_Item *mmi;
mmi = ll->data;
if (mmi->submenu == m)
mi = mmi;
}
}
/* Moving into a submenu: */
else
{
mm = l->next->data;
if (mm->entries)
mi = mm->entries->data;
}
if (mi)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/* Unselect old and select new item: */
e_menu_item_unselect(m->selected);
e_menu_item_select(mi);
e_menu_hide_submenus(mi->menu);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/* If new entry is a submenu, display it: */
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
if (mi->submenu && mi->submenu->entries)
{
e_menu_move_to(mi->submenu,
mi->menu->current.x +
mi->menu->current.w,
mi->menu->current.y + mi->y -
mi->menu->border.t);
e_menu_show(mi->submenu);
}
e_menu_update_visibility(mm);
}
}
done = 1;
}
e_menu_update_visibility(m);
}
}
/* If opened a new submenu, position it and display it: */
if (!done)
{
if (open_menus)
{
E_Menu *m;
E_Menu_Item *mi;
m = open_menus->data;
mi = m->entries->data;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
e_menu_item_select(mi);
if (mi->submenu && mi->submenu->entries)
{
e_menu_move_to(mi->submenu,
mi->menu->current.x + mi->menu->current.w,
mi->menu->current.y + mi->y -
mi->menu->border.t);
e_menu_show(mi->submenu);
}
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_init(void)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
ecore_window_get_geometry(0, NULL, NULL, &screen_w, &screen_h);
ecore_event_filter_handler_add(ECORE_EVENT_MOUSE_DOWN, e_mouse_down);
ecore_event_filter_handler_add(ECORE_EVENT_MOUSE_UP, e_mouse_up);
ecore_event_filter_handler_add(ECORE_EVENT_MOUSE_MOVE, e_mouse_move);
ecore_event_filter_handler_add(ECORE_EVENT_MOUSE_IN, e_mouse_in);
ecore_event_filter_handler_add(ECORE_EVENT_MOUSE_OUT, e_mouse_out);
ecore_event_filter_handler_add(ECORE_EVENT_WINDOW_EXPOSE, e_window_expose);
ecore_event_filter_handler_add(ECORE_EVENT_KEY_DOWN, e_key_down);
ecore_event_filter_handler_add(ECORE_EVENT_KEY_UP, e_key_up);
ecore_event_filter_handler_add(ECORE_EVENT_MOUSE_WHEEL, e_wheel);
ecore_event_filter_idle_handler_add(e_idle, NULL);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_event_win_show(void)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
/* create it */
if (!menu_event_win)
{
menu_event_win = ecore_window_input_new(0, 0, 0, screen_w, screen_h);
ecore_window_set_events(menu_event_win,
XEV_MOUSE_MOVE | XEV_BUTTON | XEV_IN_OUT |
XEV_KEY);
ecore_window_show(menu_event_win);
ecore_keyboard_grab(menu_event_win);
ecore_grab_mouse(menu_event_win, 1, 0);
}
/* raise it */
if (menu_event_win)
ecore_window_raise(menu_event_win);
if ((!ecore_grab_window_get()) || (!ecore_keyboard_grab_window_get()))
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (m->current.visible)
e_menu_hide(m);
}
e_menu_event_win_hide();
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_event_win_hide(void)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
/* destroy it */
if (menu_event_win)
{
ecore_ungrab_mouse();
ecore_keyboard_ungrab();
ecore_window_destroy(menu_event_win);
menu_event_win = 0;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/**
* e_menu_set_background - Sets the background of menu @m
* Sets background of menu using the default theme background,
* base.bits.db
*
* @m: Menu to set background on.
*/
void
e_menu_set_background(E_Menu * m)
{
char *menus;
char buf[PATH_MAX];
char *part;
int pl, pr, pt, pb;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
menus = e_config_get("menus");
part = "base.bits.db";
snprintf(buf, PATH_MAX, "%s%s", menus, part);
if ((m->bg_file) && (!strcmp(m->bg_file, buf)))
D_RETURN;
IF_FREE(m->bg_file);
m->bg_file = strdup(buf);
if (m->bg)
ebits_free(m->bg);
m->bg = ebits_load(m->bg_file);
if (m->bg)
ebits_set_color_class(m->bg, "Menu BG", 100, 200, 255, 255);
pl = pr = pt = pb = 0;
if (m->bg)
{
ebits_get_insets(m->bg, &pl, &pr, &pt, &pb);
ebits_add_to_evas(m->bg, m->evas);
ebits_move(m->bg, 0, 0);
ebits_show(m->bg);
ebits_set_layer(m->bg, 0);
}
m->current.w += ((pl + pr) - (m->border.l + m->border.r));
m->current.h += ((pt + pb) - (m->border.t + m->border.b));
m->border.l = pl;
m->border.r = pr;
m->border.t = pt;
m->border.b = pb;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_set_sel(E_Menu * m, E_Menu_Item * mi)
{
char *menus;
char buf[PATH_MAX];
int pl, pr, pt, pb;
int has_sub = 0;
int selected = 0;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
menus = e_config_get("menus");
if (!mi->separator)
{
selected = mi->selected;
if (mi->submenu)
has_sub = 1;
snprintf(buf, PATH_MAX, "%sselected-%i.submenu-%i.bits.db", menus,
selected, has_sub);
if ((mi->bg_file) && (!strcmp(mi->bg_file, buf)))
D_RETURN;
}
IF_FREE(mi->bg_file);
if (!mi->separator)
mi->bg_file = strdup(buf);
else
mi->bg_file = NULL;
if (mi->bg)
ebits_free(mi->bg);
if (mi->bg_file)
mi->bg = ebits_load(mi->bg_file);
if (mi->bg)
ebits_set_color_class(mi->bg, "Menu BG", 100, 200, 255, 255);
pl = pr = pt = pb = 0;
if (mi->bg)
{
ebits_get_insets(mi->bg, &pl, &pr, &pt, &pb);
ebits_add_to_evas(mi->bg, m->evas);
ebits_set_layer(mi->bg, 1);
}
if (m->sel_border.l < pl)
{
m->sel_border.l = pl;
m->recalc_entries = 1;
}
if (m->sel_border.r < pr)
{
m->sel_border.r = pr;
m->recalc_entries = 1;
}
if (m->sel_border.t < pt)
{
m->sel_border.t = pt;
m->recalc_entries = 1;
}
if (m->sel_border.b < pb)
{
m->sel_border.b = pb;
m->recalc_entries = 1;
}
m->redo_sel = 1;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_set_sep(E_Menu * m, E_Menu_Item * mi)
{
char *menus;
char buf[PATH_MAX];
int pl, pr, pt, pb, minx, miny;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
menus = e_config_get("menus");
snprintf(buf, PATH_MAX, "%sseparator.bits.db", menus);
if ((mi->sep_file) && (!strcmp(mi->sep_file, buf)))
D_RETURN;
IF_FREE(mi->sep_file);
mi->sep_file = strdup(buf);
if (mi->sep)
ebits_free(mi->sep);
mi->sep = ebits_load(mi->sep_file);
if (mi->sep)
ebits_set_color_class(mi->sep, "Menu BG", 100, 200, 255, 255);
pl = pr = pt = pb = 0;
minx = 0;
miny = 0;
if (mi->sep)
{
ebits_get_insets(mi->sep, &pl, &pr, &pt, &pb);
ebits_add_to_evas(mi->sep, m->evas);
ebits_set_layer(mi->sep, 1);
ebits_get_min_size(mi->sep, &minx, &miny);
}
if (mi->size.min.w < minx)
mi->size.min.w = minx;
if (mi->size.min.h < miny)
mi->size.min.h = miny;
m->redo_sel = 1;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_set_state(E_Menu * m, E_Menu_Item * mi)
{
char *menus;
char buf[PATH_MAX];
int on;
int pl, pr, pt, pb, minx, miny;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
menus = e_config_get("menus");
on = mi->on;
if (mi->check)
snprintf(buf, PATH_MAX, "%scheck-%i.bits.db", menus, on);
else
snprintf(buf, PATH_MAX, "%sradio-%i.bits.db", menus, on);
if ((mi->state_file) && (!strcmp(mi->state_file, buf)))
D_RETURN;
IF_FREE(mi->state_file);
mi->state_file = strdup(buf);
if (mi->state)
ebits_free(mi->state);
mi->state = ebits_load(mi->state_file);
if (mi->state)
ebits_set_color_class(mi->state, "Menu BG", 100, 200, 255, 255);
pl = pr = pt = pb = 0;
minx = 0;
miny = 0;
if (mi->state)
{
ebits_get_insets(mi->state, &pl, &pr, &pt, &pb);
ebits_add_to_evas(mi->state, m->evas);
ebits_set_layer(mi->state, 2);
ebits_get_min_size(mi->state, &minx, &miny);
}
if (mi->size.min.w < minx)
mi->size.min.w = minx;
if (mi->size.min.h < miny)
mi->size.min.h = miny;
m->redo_sel = 1;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
static void
e_menu_cleanup(E_Menu * m)
2000-12-29 12:49:17 -08:00
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
for (l = m->entries; l; l = l->next)
{
E_Menu_Item *mi;
mi = l->data;
e_menu_item_unrealize(m, mi);
IF_FREE(mi->str);
IF_FREE(mi->icon);
free(mi);
}
m->entries = evas_list_free(m->entries);
m->selected = NULL;
IF_FREE(m->bg_file);
evas_free(m->evas);
ecore_window_destroy(m->win.main);
menus = evas_list_remove(menus, m);
open_menus = evas_list_remove(open_menus, m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
/* Call the destructor of the base class */
e_object_cleanup(E_OBJECT(m));
m = NULL;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
E_Menu *
e_menu_new(void)
{
E_Menu *m;
char *font_dir;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
font_dir = e_config_get("fonts");
m = NEW(E_Menu, 1);
ZERO(m, E_Menu, 1);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
e_object_init(E_OBJECT(m), (E_Cleanup_Func) e_menu_cleanup);
m->win.main = ecore_window_override_new(0, 0, 0, 1, 1);
2003-01-15 10:29:56 -08:00
m->evas = e_evas_new_all(ecore_display_get(),
m->win.main,
2003-01-15 10:29:56 -08:00
0, 0, 1, 1, font_dir);
/* aaaaaaaaah. this makes building the menu fast - moves the mouse far */
/* far far far far away so callbacks and events arent triggerd as we */
/* create objects that ofter hang around 0,0 - the default place for */
/* the pointer to be... this means my 2000 entry menu works and comes up */
/* pretty damn fast - considering i creating it when i click :) - problem */
/* you can't fit 2000 entires into a window in X - since the limit is */
/* 65536x65536 fo X - the other problem is i can only really draw in */
/* the first 32768x32768 pixels of the window - thus limiting the menu */
/* size to well - 32768x32768 - normally ok - but in extremes not so */
/* good. We *COULD* do a workaround that meant we did the menu scrolling */
/* within the evas and faked a menu window that only gets as big as the */
/* screen - an then re-render it all - but well.. it's an extreme and */
/* for now i think people will just have to live with a maximum menu size */
/* of 32768x32768... didums! */
2003-01-15 10:29:56 -08:00
evas_event_feed_mouse_move(m->evas, -999999999, -99999999);
2003-01-15 10:29:56 -08:00
m->win.evas = e_evas_get_window(m->evas);
ecore_window_set_events(m->win.evas,
XEV_EXPOSE | XEV_MOUSE_MOVE | XEV_BUTTON | XEV_IN_OUT
| XEV_KEY);
ecore_window_set_events(m->win.main, XEV_IN_OUT | XEV_KEY);
ecore_window_show(m->win.evas);
e_menu_set_background(m);
m->current.w = m->border.l + m->border.r;
m->current.h = m->border.t + m->border.b;
m->changed = 1;
menus = evas_list_prepend(menus, m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN_(m);
}
2000-12-29 12:49:17 -08:00
void
e_menu_hide(E_Menu * m)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (m->selected)
{
m->selected->selected = 0;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
}
m->selected = NULL;
m->redo_sel = 1;
m->current.visible = 0;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_show(E_Menu * m)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
m->current.visible = 1;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_move_to(E_Menu * m, int x, int y)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
m->current.x = x;
m->current.y = y;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_show_at_mouse(E_Menu * m, int x, int y, Time t)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
D("show at mouse\n");
m->current.x = x;
m->current.y = y;
m->time = t;
D("show menu\n") e_menu_show(m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_add_item(E_Menu * m, E_Menu_Item * mi)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
m->entries = evas_list_append(m->entries, mi);
m->recalc_entries = 1;
m->changed = 1;
mi->menu = m;
e_menu_item_realize(m, mi);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_del_item(E_Menu * m, E_Menu_Item * mi)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
m->entries = evas_list_remove(m->entries, mi);
m->recalc_entries = 1;
m->changed = 1;
e_menu_item_unrealize(m, mi);
IF_FREE(mi->str);
IF_FREE(mi->icon);
if (mi->menu->selected == mi)
mi->menu->selected = NULL;
2002-03-11 12:03:14 -08:00
FREE(mi);
mi->menu = NULL;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_update(E_Menu * m, E_Menu_Item * mi)
{
int tx, ty, tw, th, ix, iy, iw, ih, rx, ry, rw, rh;
double dtw, dth;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (mi->sep)
{
ebits_move(mi->sep, mi->x, mi->y);
ebits_resize(mi->sep, mi->size.w + m->sel_border.l + m->sel_border.r,
mi->size.h);
ebits_show(mi->sep);
}
else
{
rx = 0;
ry = 0;
rh = 0;
rw = m->size.state;
if (mi->state)
{
ebits_get_min_size(mi->state, &rw, &rh);
rx = 0;
ry = ((mi->size.h - rh) / 2);
ebits_move(mi->state, m->sel_border.l + mi->x + rx,
m->sel_border.t + mi->y + ry);
ebits_resize(mi->state, rw, rh);
}
tx = 0;
ty = 0;
tw = 0;
th = 0;
if (mi->obj_text)
{
2003-01-15 10:29:56 -08:00
evas_object_geometry_get(mi->obj_text, NULL, NULL, &dtw, &dth);
tw = (int)dtw;
th = (int)dth;
}
ix = 0;
iy = 0;
iw = 0;
ih = 0;
if (mi->obj_icon)
{
int sh;
2003-01-15 10:29:56 -08:00
evas_object_image_size_get(mi->obj_icon, &iw, &ih);
sh = th;
if (rh > th)
sh = rh;
if ((mi->scale_icon) && (ih > sh) && (mi->str))
{
iw = (iw * sh) / ih;
ih = sh;
}
if (m->size.state)
ix = rx + m->size.state + m->pad.state;
ix += ((m->size.icon - iw) / 2);
iy = ((mi->size.h - ih) / 2);
2003-01-15 10:29:56 -08:00
evas_object_move(mi->obj_icon, m->sel_border.l + mi->x + ix,
m->sel_border.t + mi->y + iy);
2003-01-15 10:29:56 -08:00
evas_object_resize(mi->obj_icon, iw, ih);
evas_object_image_fill_set(mi->obj_icon, 0, 0, iw, ih);
}
if (mi->obj_text)
{
if (m->size.state)
tx = rx + m->size.state + m->pad.state;
if (m->size.icon)
tx += m->size.icon + m->pad.icon;
ty = ((mi->size.h - th) / 2);
2003-01-15 10:29:56 -08:00
evas_object_move(mi->obj_text, m->sel_border.l + mi->x + tx,
m->sel_border.t + mi->y + ty);
}
if (mi->obj_entry)
{
2003-01-15 10:29:56 -08:00
evas_object_move(mi->obj_entry, mi->x, mi->y);
evas_object_resize(mi->obj_entry,
mi->size.w + m->sel_border.l + m->sel_border.r,
mi->size.h + m->sel_border.t + m->sel_border.b);
}
if (mi->state)
{
ebits_show(mi->state);
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_unrealize(E_Menu * m, E_Menu_Item * mi)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (mi->bg)
ebits_free(mi->bg);
mi->bg = NULL;
IF_FREE(mi->bg_file);
mi->bg_file = NULL;
if (mi->obj_entry)
2003-01-15 10:29:56 -08:00
evas_object_del(mi->obj_entry);
mi->obj_entry = NULL;
if (mi->obj_text)
2003-01-15 10:29:56 -08:00
evas_object_del(mi->obj_text);
mi->obj_text = NULL;
if (mi->obj_icon)
2003-01-15 10:29:56 -08:00
evas_object_del(mi->obj_icon);
mi->obj_icon = NULL;
if (mi->state)
ebits_free(mi->state);
mi->state = NULL;
IF_FREE(mi->state_file);
mi->state_file = NULL;
if (mi->sep)
ebits_free(mi->sep);
mi->sep = NULL;
IF_FREE(mi->sep_file);
mi->sep_file = NULL;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2003-01-15 10:29:56 -08:00
UN(m);
}
void
e_menu_item_realize(E_Menu * m, E_Menu_Item * mi)
{
double tw, th;
int iw, ih, rw, rh;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (mi->separator)
{
e_menu_set_sep(m, mi);
}
else
{
if (mi->str)
{
2003-01-15 10:29:56 -08:00
mi->obj_text = evas_object_text_add(m->evas);
evas_object_text_font_set(mi->obj_text, "menu", 8);
2003-01-15 10:29:56 -08:00
evas_object_text_text_set(mi->obj_text, mi->str);
evas_object_color_set(mi->obj_text, 0, 0, 0, 255);
evas_object_show(mi->obj_text);
evas_object_layer_set(mi->obj_text, 10);
}
if (mi->icon)
{
2003-01-15 10:29:56 -08:00
mi->obj_icon = evas_object_image_add(m->evas);
evas_object_image_file_set(mi->obj_icon, mi->icon, NULL);
evas_object_show(mi->obj_icon);
evas_object_layer_set(mi->obj_icon, 10);
}
2003-01-15 10:29:56 -08:00
mi->obj_entry = evas_object_rectangle_add(m->evas);
evas_object_layer_set(mi->obj_entry, 11);
evas_object_color_set(mi->obj_entry, 0, 0, 0, 0);
evas_object_show(mi->obj_entry);
tw = 0;
th = 0;
if (mi->obj_text)
2003-01-15 10:29:56 -08:00
evas_object_geometry_get(mi->obj_text, NULL, NULL, &tw, &th);
iw = 0;
ih = 0;
if (mi->obj_icon)
2003-01-15 10:29:56 -08:00
evas_object_image_size_get(mi->obj_icon, &iw, &ih);
rw = 0;
rh = 0;
if (mi->state)
ebits_get_min_size(mi->state, &rw, &rh);
mi->size.min.w = (int)tw + rw;
if (rh > th)
th = (double)rh;
if (((!mi->scale_icon) && (ih > th)) || ((!mi->str) && (ih > th)))
th = (double)ih;
mi->size.min.h = (int)th;
2003-01-15 10:29:56 -08:00
/* checkme todo */
evas_object_event_callback_add(mi->obj_entry, EVAS_CALLBACK_MOUSE_IN,
e_menu_item_in_cb, mi);
2003-01-15 10:29:56 -08:00
evas_object_event_callback_add(mi->obj_entry, EVAS_CALLBACK_MOUSE_OUT,
e_menu_item_out_cb, mi);
e_menu_set_sel(m, mi);
if ((mi->radio) || (mi->check))
e_menu_set_state(m, mi);
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
E_Menu_Item *
e_menu_item_new(char *str)
{
E_Menu_Item *mi;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi = NEW(E_Menu_Item, 1);
ZERO(mi, E_Menu_Item, 1);
if (str)
mi->str = strdup(str);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN_(mi);
}
void
e_menu_obscure_outside_screen(E_Menu * m)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
/* obscure stuff outside the screen boundaries - optimizes rendering */
2003-01-15 10:29:56 -08:00
evas_obscured_clear(m->evas);
evas_obscured_rectangle_add(m->evas,
-m->current.x - 100000,
-m->current.y - 100000, 200000 + screen_w, 100000);
2003-01-15 10:29:56 -08:00
evas_obscured_rectangle_add(m->evas,
-m->current.x - 100000,
-m->current.y - 100000, 100000, 200000 + screen_h);
2003-01-15 10:29:56 -08:00
evas_obscured_rectangle_add(m->evas,
-m->current.x - 100000,
screen_h - m->current.y, 200000 + screen_w, 100000);
2003-01-15 10:29:56 -08:00
evas_obscured_rectangle_add(m->evas,
screen_w - m->current.x,
-m->current.y - 100000, 100000, 200000 + screen_h);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_scroll_all_by(int dx, int dy)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
for (l = menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
if (m->current.visible)
{
m->current.x += dx;
m->current.y += dy;
m->changed = 1;
}
}
if (!keyboard_nav)
{
for (l = open_menus; l; l = l->next)
{
E_Menu *m;
m = l->data;
2003-01-15 10:29:56 -08:00
evas_event_feed_mouse_move(m->evas,
mouse_x - m->current.x, mouse_y - m->current.y);
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_update_visibility(E_Menu * m)
{
E_Menu_Item *mi;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi = m->selected;
if (mi)
{
/* if the entry is off screen - scroll so it's on screen */
if (m->current.x < 0)
e_menu_scroll_all_by(-m->current.x, 0);
else if ((m->current.x + m->current.w) > screen_w)
{
e_menu_scroll_all_by(screen_w - (m->current.x + m->current.w), 0);
}
if ((m->current.y + m->sel_border.t + mi->y) < 0)
{
if (m->current.y < -(screen_h / 4))
e_menu_scroll_all_by(0, screen_h / 4);
else
e_menu_scroll_all_by(0, -m->current.y);
}
else if ((m->current.y + m->sel_border.t + mi->y + mi->size.h) >
screen_h)
{
if ((m->current.y + m->current.h - screen_h) < (screen_h / 4))
e_menu_scroll_all_by(0,
-(m->current.y + m->current.h - screen_h));
else
e_menu_scroll_all_by(0, -(screen_h / 4));
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_update_base(E_Menu * m)
{
int size_changed = 0;
int location_changed = 0;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (!m->changed)
D_RETURN;
if (m->recalc_entries)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
int max_w, max_h;
int i;
max_w = 0;
max_h = 0;
for (l = m->entries; l; l = l->next)
{
E_Menu_Item *mi;
mi = l->data;
if (mi->size.min.h > max_h)
max_h = mi->size.min.h;
}
m->size.state = 0;
m->size.icon = 0;
m->size.text = 0;
for (l = m->entries; l; l = l->next)
{
E_Menu_Item *mi;
int iw, ih, rw, rh;
double tw, th;
mi = l->data;
if (!mi->separator)
{
tw = 0;
th = 0;
if (mi->obj_text)
2003-01-15 10:29:56 -08:00
evas_object_geometry_get(mi->obj_text, NULL, NULL, &tw,
&th);
iw = 0;
ih = 0;
if (mi->obj_icon)
2003-01-15 10:29:56 -08:00
evas_object_image_size_get(mi->obj_icon, &iw, &ih);
rw = 0;
rh = 0;
if (mi->state)
ebits_get_min_size(mi->state, &rw, &rh);
if (m->size.text < tw)
m->size.text = tw;
if (m->size.state < rw)
m->size.state = rw;
if ((mi->scale_icon) && (iw > 0) && (ih > 0) && (mi->str))
{
int iiw;
iiw = iw;
if (ih > (int)th)
iiw = (iw * (int)th) / ih;
if (m->size.icon < iiw)
m->size.icon = iiw;
}
else if (m->size.icon < iw)
m->size.icon = iw;
}
}
max_w = m->size.state;
if (m->size.state)
max_w += m->pad.state;
max_w += m->size.icon;
if (m->size.icon)
max_w += m->pad.icon;
max_w += m->size.text;
i = m->border.t;
for (l = m->entries; l; l = l->next)
{
E_Menu_Item *mi;
mi = l->data;
mi->size.w = max_w;
if (mi->separator)
mi->size.h = mi->size.min.h;
else
mi->size.h = max_h;
mi->x = m->border.l;
mi->y = i;
if (!mi->separator)
i += m->sel_border.t + m->sel_border.b;
if (mi->separator)
i += mi->size.h;
else
i += max_h;
e_menu_item_update(m, mi);
}
m->current.w =
m->border.l + m->border.r + max_w + m->sel_border.l +
m->sel_border.r;
m->current.h = m->border.b + i;
m->recalc_entries = 0;
}
if (m->redo_sel)
{
2003-01-15 10:29:56 -08:00
Evas_List * l;
for (l = m->entries; l; l = l->next)
{
E_Menu_Item *mi;
mi = l->data;
e_menu_set_sel(m, mi);
if (mi)
{
if (mi->bg)
{
ebits_move(mi->bg, mi->x, mi->y);
ebits_resize(mi->bg,
mi->size.w + m->sel_border.l +
m->sel_border.r,
mi->size.h + m->sel_border.t +
m->sel_border.b);
ebits_show(mi->bg);
}
}
}
m->redo_sel = 0;
}
if ((m->current.x != m->previous.x) || (m->current.y != m->previous.y))
location_changed = 1;
if ((m->current.w != m->previous.w) || (m->current.h != m->previous.h))
size_changed = 1;
if ((location_changed) && (size_changed))
{
ecore_window_move_resize(m->win.main, m->current.x, m->current.y,
m->current.w, m->current.h);
e_menu_obscure_outside_screen(m);
}
else if (location_changed)
{
ecore_window_move(m->win.main, m->current.x, m->current.y);
e_menu_obscure_outside_screen(m);
}
else if (size_changed)
{
ecore_window_resize(m->win.main, m->current.w, m->current.h);
}
if (size_changed)
{
ecore_window_resize(m->win.evas, m->current.w, m->current.h);
2003-01-15 10:29:56 -08:00
evas_output_size_set(m->evas, m->current.w, m->current.h);
evas_output_viewport_set(m->evas, 0, 0, m->current.w, m->current.h);
if (m->bg)
ebits_resize(m->bg, m->current.w, m->current.h);
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
2003-01-15 10:29:56 -08:00
e_menu_update_shape(m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_update_finish(E_Menu * m)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (!m->changed)
D_RETURN;
m->previous = m->current;
m->changed = 0;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_update_shows(E_Menu * m)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (!m->changed)
D_RETURN;
if (m->current.visible != m->previous.visible)
{
if (m->current.visible)
{
ecore_window_raise(m->win.main);
e_menu_event_win_show();
ecore_window_show(m->win.main);
if (!open_menus)
keyboard_nav = 0;
open_menus = evas_list_append(open_menus, m);
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
2003-01-15 10:29:56 -08:00
e_menu_update_shape(m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_update_hides(E_Menu * m)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (!m->changed)
D_RETURN;
if (m->current.visible != m->previous.visible)
{
if (!m->current.visible)
{
if (m->selected)
{
E_Menu_Item *mi;
mi = m->selected;
mi->selected = 0;
e_menu_set_sel(m, mi);
if (mi)
{
if (mi->bg)
{
ebits_move(mi->bg, mi->x, mi->y);
ebits_resize(mi->bg,
mi->size.w + m->sel_border.l +
m->sel_border.r,
mi->size.h + m->sel_border.t +
m->sel_border.b);
ebits_show(mi->bg);
}
}
m->redo_sel = 1;
m->changed = 1;
m->selected = NULL;
}
open_menus = evas_list_remove(open_menus, m);
ecore_window_hide(m->win.main);
if (!open_menus)
e_menu_event_win_hide();
if (m->func_hide)
m->func_hide(m, m->func_hide_data);
}
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
2003-01-15 10:29:56 -08:00
e_menu_update_shape(m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
2003-01-15 10:29:56 -08:00
void
e_menu_update_shape(E_Menu *m)
{
Pixmap pmap, mask;
D_ENTER;
pmap = ecore_pixmap_new(m->win.main, m->current.w, m->current.h, 0);
mask = ecore_pixmap_new(m->win.main, m->current.w, m->current.h, 1);
e_evas_get_mask(m->evas, pmap, mask);
ecore_window_set_background_pixmap(m->win.main, pmap);
ecore_window_set_shape_mask(m->win.main, mask);
ecore_window_clear(m->win.main);
ecore_pixmap_free(pmap);
ecore_pixmap_free(mask);
D_RETURN;
}
2000-12-29 12:49:17 -08:00
void
e_menu_update(E_Menu * m)
2000-12-29 12:49:17 -08:00
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
e_menu_update_base(m);
e_menu_update_shows(m);
e_menu_update_hides(m);
e_menu_update_finish(m);
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
2000-12-29 12:49:17 -08:00
}
void
e_menu_item_set_icon(E_Menu_Item * mi, char *icon)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
IF_FREE(mi->icon);
mi->icon = NULL;
if (icon)
mi->icon = strdup(icon);
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_text(E_Menu_Item * mi, char *text)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
IF_FREE(mi->str);
mi->str = NULL;
if (text)
mi->str = strdup(text);
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_separator(E_Menu_Item * mi, int sep)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi->separator = sep;
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_radio(E_Menu_Item * mi, int radio)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi->radio = radio;
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_check(E_Menu_Item * mi, int check)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi->check = check;
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_state(E_Menu_Item * mi, int state)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi->on = state;
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->redo_sel = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_submenu(E_Menu_Item * mi, E_Menu * submenu)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
if (mi->submenu)
e_menu_hide(mi->submenu);
mi->submenu = submenu;
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->redo_sel = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_item_set_scale_icon(E_Menu_Item * mi, int scale)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
mi->scale_icon = scale;
if (mi->menu)
{
mi->menu->recalc_entries = 1;
mi->menu->changed = 1;
}
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_set_padding_icon(E_Menu * m, int pad)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
m->pad.icon = pad;
m->recalc_entries = 1;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}
void
e_menu_set_padding_state(E_Menu * m, int pad)
{
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_ENTER;
m->pad.state = pad;
m->recalc_entries = 1;
m->changed = 1;
Alright, I spent some time now reading e17's code. Here's what I've changed, this is big, so read this carefully :) * I've added debugging macros for messages and function call tracing. Usage: D("Creating item %i %i %i\n", x, y, z); Define DEBUG to use the D macro. D_ENTER; D_RETURN; D_RETURN_(x); These are for call tracing. Use D_RETURN_(x) when returning something from a function. Define DEBUG_NEST to use this. * added iconbar header file to Makefile.am * added proper new()/cleanup() calls for E_Delayed_Action; * I've completely rewritten the object and observer handling. Bye bye macros, this was nasty. It'll be hard enough to avoid leaks with usecounting in C. We now basically have the same system as gtk. There's a clear separation of observer and object code now. An E_Object by itself has nothing to do with observing or being observed, therefore, there are now E_Observers and E_Observees that are derived from E_Object. IMPORTANT: The cleanup system now reflects the reference count system, therefore, all ..._free() calls are now static, because the destructor should never be called explicitly, but implicitly through e_object_unref(). The object handling now is as follows: - The cleanup functions clean up everything that is contained in a struct, but NOT the struct itself. Instead of the final free() call, they call the destructor of the base class. The calls will walk up the hierarchy and clean up what's contained in every struct, and the final e_object_cleanup() will free the structure itself. E_Delayed_Action is a good example. - The only calls that influence the reference count are e_object_ref() and e_object_unref(). If you need to do things before an object gets destroyed, you can query the use count using e_object_get_usecount() and check if it's equal to 1. So this: OBJ_UNREF(b); OBJ_IF_FREE(b) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); OBJ_FREE(b); } now is this: if (e_object_get_usecount(E_OBJECT(b)) == 1) { ecore_window_reparent(e->win, 0, 0, 0); e_icccm_release(e->win); } e_object_unref(E_OBJECT(b)); object.h and observer.h are completely commented, it shouldn't be too hard to understand. This'll need to be documented in the manual anyway. * E_Objects are now used in lots of places where void* were used as pointers to objects before, especially in the actions code. This is obviously better, as it will generate compiler warnings when people want to pass things to functions that expect E_Objects. This could probably be more restrictive. * Added typedefs for the function prototypes in E_Action_Impl. Those fat signatures were just painful to read in the function declarations/implementations. * I've also tried to give parameters more useful names. Calling an object "o" is a lot of fun when you want to grep for it. * Included is also Graham's latest menu.c patch. Sorry for the delay, Graham. * I've added checks to the menu code that make sure that menus don't pop up when they're empty (which resulted in a little useless rectangle). I guess that's it for now. Sorry if I broke anything, but this was necessary imho. SVN revision: 5605
2001-11-02 09:07:52 -08:00
D_RETURN;
}