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"
|
2000-12-08 14:54:42 -08:00
|
|
|
#include "e.h"
|
2001-07-30 09:59:37 -07:00
|
|
|
#include "border.h"
|
2001-07-30 18:12:02 -07:00
|
|
|
#include "icccm.h"
|
2001-08-16 01:45:37 -07:00
|
|
|
#include "util.h"
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
|
|
|
|
|
/* Motif window hints */
|
|
|
|
|
#define MWM_HINTS_FUNCTIONS (1L << 0)
|
|
|
|
|
#define MWM_HINTS_DECORATIONS (1L << 1)
|
|
|
|
|
#define MWM_HINTS_INPUT_MODE (1L << 2)
|
|
|
|
|
#define MWM_HINTS_STATUS (1L << 3)
|
|
|
|
|
|
|
|
|
|
/* bit definitions for MwmHints.functions */
|
|
|
|
|
#define MWM_FUNC_ALL (1L << 0)
|
|
|
|
|
#define MWM_FUNC_RESIZE (1L << 1)
|
|
|
|
|
#define MWM_FUNC_MOVE (1L << 2)
|
|
|
|
|
#define MWM_FUNC_MINIMIZE (1L << 3)
|
|
|
|
|
#define MWM_FUNC_MAXIMIZE (1L << 4)
|
|
|
|
|
#define MWM_FUNC_CLOSE (1L << 5)
|
|
|
|
|
|
|
|
|
|
/* bit definitions for MwmHints.decorations */
|
|
|
|
|
#define MWM_DECOR_ALL (1L << 0)
|
|
|
|
|
#define MWM_DECOR_BORDER (1L << 1)
|
|
|
|
|
#define MWM_DECOR_RESIZEH (1L << 2)
|
|
|
|
|
#define MWM_DECOR_TITLE (1L << 3)
|
|
|
|
|
#define MWM_DECOR_MENU (1L << 4)
|
|
|
|
|
#define MWM_DECOR_MINIMIZE (1L << 5)
|
|
|
|
|
#define MWM_DECOR_MAXIMIZE (1L << 6)
|
|
|
|
|
|
|
|
|
|
/* bit definitions for MwmHints.inputMode */
|
|
|
|
|
#define MWM_INPUT_MODELESS 0
|
|
|
|
|
#define MWM_INPUT_PRIMARY_APPLICATION_MODAL 1
|
|
|
|
|
#define MWM_INPUT_SYSTEM_MODAL 2
|
|
|
|
|
#define MWM_INPUT_FULL_APPLICATION_MODAL 3
|
|
|
|
|
|
|
|
|
|
#define PROP_MWM_HINTS_ELEMENTS 5
|
|
|
|
|
|
|
|
|
|
/* Motif window hints */
|
|
|
|
|
typedef struct _mwmhints
|
|
|
|
|
{
|
|
|
|
|
int flags;
|
|
|
|
|
int functions;
|
|
|
|
|
int decorations;
|
|
|
|
|
int inputMode;
|
|
|
|
|
int status;
|
|
|
|
|
}
|
|
|
|
|
MWMHints;
|
2000-12-08 14:54:42 -08:00
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_move_resize(Window win, int x, int y, int w, int 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_ENTER;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_send_event_move_resize(win, x, y, w, 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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
2001-11-15 21:39:34 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_send_focus_to(Window win, int takes_focus)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_wm_take_focus = 0;
|
|
|
|
|
static Atom a_wm_protocols = 0;
|
|
|
|
|
int msg_focus = 0;
|
|
|
|
|
int *props;
|
|
|
|
|
int size;
|
|
|
|
|
|
|
|
|
|
D_ENTER;
|
|
|
|
|
|
|
|
|
|
ECORE_ATOM(a_wm_take_focus, "WM_TAKE_FOCUS");
|
|
|
|
|
ECORE_ATOM(a_wm_protocols, "WM_PROTOCOLS");
|
|
|
|
|
|
|
|
|
|
props = ecore_window_property_get(win, a_wm_protocols, XA_ATOM, &size);
|
|
|
|
|
if (props)
|
|
|
|
|
{
|
|
|
|
|
int i, num;
|
|
|
|
|
|
|
|
|
|
num = size / sizeof(int);
|
|
|
|
|
for (i = 0; i < num; i++)
|
|
|
|
|
{
|
|
|
|
|
if (props[i] == (int)a_wm_take_focus) msg_focus = 1;
|
|
|
|
|
}
|
|
|
|
|
FREE(props);
|
|
|
|
|
}
|
|
|
|
|
if (takes_focus)
|
|
|
|
|
ecore_focus_to_window(win);
|
|
|
|
|
if (msg_focus)
|
|
|
|
|
{
|
|
|
|
|
unsigned int data[5];
|
|
|
|
|
|
|
|
|
|
data[0] = a_wm_take_focus;
|
|
|
|
|
data[1] = CurrentTime;
|
|
|
|
|
ecore_window_send_client_message(win, a_wm_protocols, 32, data);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
D_RETURN;
|
|
|
|
|
}
|
|
|
|
|
|
2000-12-08 14:54:42 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_delete(Window win)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_wm_delete_window = 0;
|
|
|
|
|
static Atom a_wm_protocols = 0;
|
|
|
|
|
int *props;
|
|
|
|
|
int size;
|
|
|
|
|
int del_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_ENTER;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_wm_delete_window, "WM_DELETE_WINDOW");
|
|
|
|
|
ECORE_ATOM(a_wm_protocols, "WM_PROTOCOLS");
|
2000-12-08 14:54:42 -08:00
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
props = ecore_window_property_get(win, a_wm_protocols, XA_ATOM, &size);
|
2000-12-08 14:54:42 -08:00
|
|
|
if (props)
|
|
|
|
|
{
|
|
|
|
|
int i, num;
|
|
|
|
|
|
|
|
|
|
num = size / sizeof(int);
|
|
|
|
|
for (i = 0; i < num; i++)
|
|
|
|
|
{
|
|
|
|
|
if (props[i] == (int)a_wm_delete_window) del_win = 1;
|
|
|
|
|
}
|
|
|
|
|
FREE(props);
|
|
|
|
|
}
|
|
|
|
|
if (del_win)
|
|
|
|
|
{
|
|
|
|
|
unsigned int data[5];
|
|
|
|
|
|
|
|
|
|
data[0] = a_wm_delete_window;
|
|
|
|
|
data[1] = CurrentTime;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_send_client_message(win, a_wm_protocols, 32, data);
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_kill_client(win);
|
2000-12-08 14:54:42 -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_RETURN;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_state_mapped(Window win)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_wm_state = 0;
|
|
|
|
|
unsigned int data[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_wm_state, "WM_STATE");
|
2000-12-08 14:54:42 -08:00
|
|
|
data[0] = NormalState;
|
|
|
|
|
data[1] = 0;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_wm_state, a_wm_state, 32, data, 2);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_state_iconified(Window win)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_wm_state = 0;
|
|
|
|
|
unsigned int data[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_wm_state, "WM_STATE");
|
2000-12-08 14:54:42 -08:00
|
|
|
data[0] = IconicState;
|
|
|
|
|
data[1] = 0;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_wm_state, a_wm_state, 32, data, 2);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_state_withdrawn(Window win)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_wm_state = 0;
|
|
|
|
|
unsigned int data[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_wm_state, "WM_STATE");
|
2000-12-08 14:54:42 -08:00
|
|
|
data[0] = WithdrawnState;
|
|
|
|
|
data[1] = 0;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_wm_state, a_wm_state, 32, data, 2);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_adopt(Window 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_ENTER;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_add_to_save_set(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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_release(Window 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_ENTER;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_del_from_save_set(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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
2001-04-02 12:03:55 -07:00
|
|
|
void
|
|
|
|
|
e_icccm_get_pos_info(Window win, E_Border *b)
|
|
|
|
|
{
|
|
|
|
|
XSizeHints hint;
|
|
|
|
|
int mask;
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
if (ecore_window_get_wm_size_hints(win, &hint, &mask))
|
2001-04-02 12:03:55 -07:00
|
|
|
{
|
|
|
|
|
if ((hint.flags & USPosition) || ((hint.flags & PPosition)))
|
|
|
|
|
{
|
|
|
|
|
int x, y, w, h;
|
|
|
|
|
|
2001-11-03 01:07:40 -08:00
|
|
|
D("%li %li\n", hint.flags & USPosition, hint.flags & PPosition);
|
2001-04-02 12:03:55 -07:00
|
|
|
b->client.pos.requested = 1;
|
|
|
|
|
b->client.pos.gravity = NorthWestGravity;
|
|
|
|
|
if (hint.flags & PWinGravity)
|
|
|
|
|
b->client.pos.gravity = hint.win_gravity;
|
|
|
|
|
x = y = w = h = 0;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_get_geometry(win, &x, &y, &w, &h);
|
2001-04-02 12:03:55 -07:00
|
|
|
b->client.pos.x = x;
|
|
|
|
|
b->client.pos.y = y;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
b->client.pos.requested = 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;
|
2001-04-02 12:03:55 -07:00
|
|
|
}
|
|
|
|
|
|
2000-12-08 14:54:42 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_get_size_info(Window win, E_Border *b)
|
|
|
|
|
{
|
|
|
|
|
int base_w, base_h, min_w, min_h, max_w, max_h, grav, step_w, step_h;
|
|
|
|
|
double aspect_min, aspect_max;
|
|
|
|
|
int mask;
|
|
|
|
|
XSizeHints hint;
|
|
|
|
|
|
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;
|
|
|
|
|
|
2000-12-08 14:54:42 -08:00
|
|
|
grav = NorthWestGravity;
|
|
|
|
|
mask = 0;
|
|
|
|
|
min_w = 0;
|
|
|
|
|
min_h = 0;
|
|
|
|
|
max_w = 65535;
|
|
|
|
|
max_h = 65535;
|
|
|
|
|
aspect_min = 0.0;
|
|
|
|
|
aspect_max = 999999.0;
|
|
|
|
|
step_w = 1;
|
|
|
|
|
step_h = 1;
|
|
|
|
|
base_w = 0;
|
|
|
|
|
base_h = 0;
|
2001-10-17 15:34:27 -07:00
|
|
|
if (ecore_window_get_wm_size_hints(win, &hint, &mask))
|
2000-12-08 14:54:42 -08:00
|
|
|
{
|
|
|
|
|
if (hint.flags & PMinSize)
|
|
|
|
|
{
|
|
|
|
|
min_w = hint.min_width;
|
|
|
|
|
min_h = hint.min_height;
|
|
|
|
|
}
|
|
|
|
|
if (hint.flags & PMaxSize)
|
|
|
|
|
{
|
|
|
|
|
max_w = hint.max_width;
|
|
|
|
|
max_h = hint.max_height;
|
|
|
|
|
if (max_w < min_w) max_w = min_w;
|
|
|
|
|
if (max_h < min_h) max_h = min_h;
|
|
|
|
|
}
|
|
|
|
|
if (hint.flags & PResizeInc)
|
|
|
|
|
{
|
|
|
|
|
step_w = hint.width_inc;
|
|
|
|
|
step_h = hint.height_inc;
|
|
|
|
|
if (step_w < 1) step_w = 1;
|
|
|
|
|
if (step_h < 1) step_h = 1;
|
|
|
|
|
}
|
|
|
|
|
if (hint.flags & PBaseSize)
|
|
|
|
|
{
|
|
|
|
|
base_w = hint.base_width;
|
|
|
|
|
base_h = hint.base_height;
|
|
|
|
|
if (base_w > max_w) max_w = base_w;
|
|
|
|
|
if (base_h > max_h) max_h = base_h;
|
|
|
|
|
}
|
|
|
|
|
else
|
|
|
|
|
{
|
|
|
|
|
base_w = min_w;
|
|
|
|
|
base_h = min_h;
|
|
|
|
|
}
|
|
|
|
|
if (hint.flags & PAspect)
|
|
|
|
|
{
|
|
|
|
|
if (hint.min_aspect.y > 0)
|
|
|
|
|
aspect_min = ((double)hint.min_aspect.x) / ((double)hint.min_aspect.y);
|
|
|
|
|
if (hint.max_aspect.y > 0)
|
|
|
|
|
aspect_max = ((double)hint.max_aspect.x) / ((double)hint.max_aspect.y);
|
|
|
|
|
}
|
|
|
|
|
}
|
2001-03-22 10:10:08 -08:00
|
|
|
b->client.min.w = min_w;
|
|
|
|
|
b->client.min.h = min_h;
|
|
|
|
|
b->client.max.w = max_w;
|
|
|
|
|
b->client.max.h = max_h;
|
|
|
|
|
b->client.base.w = base_w;
|
|
|
|
|
b->client.base.h = base_h;
|
|
|
|
|
b->client.step.w = step_w;
|
|
|
|
|
b->client.step.h = step_h;
|
|
|
|
|
b->client.min.aspect = aspect_min;
|
|
|
|
|
b->client.max.aspect = aspect_max;
|
|
|
|
|
b->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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_mwm_hints(Window win, E_Border *b)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_motif_wm_hints = 0;
|
|
|
|
|
MWMHints *mwmhints;
|
|
|
|
|
int size;
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_motif_wm_hints, "_MOTIF_WM_HINTS");
|
2000-12-08 14:54:42 -08:00
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
mwmhints = ecore_window_property_get(win, a_motif_wm_hints, a_motif_wm_hints, &size);
|
2000-12-08 14:54:42 -08:00
|
|
|
if (mwmhints)
|
|
|
|
|
{
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
int num;
|
2000-12-08 14:54:42 -08:00
|
|
|
|
|
|
|
|
num = size / sizeof(int);
|
|
|
|
|
if (num < PROP_MWM_HINTS_ELEMENTS)
|
|
|
|
|
{
|
|
|
|
|
FREE(mwmhints);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
if (mwmhints->flags & MWM_HINTS_DECORATIONS)
|
|
|
|
|
{
|
|
|
|
|
b->client.border = 0;
|
|
|
|
|
b->client.handles = 0;
|
|
|
|
|
b->client.titlebar = 0;
|
|
|
|
|
if (mwmhints->decorations & MWM_DECOR_ALL)
|
|
|
|
|
{
|
|
|
|
|
b->client.border = 1;
|
|
|
|
|
b->client.handles = 1;
|
|
|
|
|
b->client.titlebar = 1;
|
|
|
|
|
}
|
|
|
|
|
if (mwmhints->decorations & MWM_DECOR_BORDER) b->client.border = 1;
|
|
|
|
|
if (mwmhints->decorations & MWM_DECOR_RESIZEH) b->client.handles = 1;
|
|
|
|
|
if (mwmhints->decorations & MWM_DECOR_TITLE) b->client.titlebar = 1;
|
2001-01-02 15:10:12 -08:00
|
|
|
e_border_apply_border(b);
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
FREE(mwmhints);
|
|
|
|
|
}
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_layer(Window win, E_Border *b)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_win_layer = 0;
|
|
|
|
|
int *props;
|
|
|
|
|
int size;
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_layer, "_WIN_LAYER");
|
2000-12-08 14:54:42 -08:00
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
props = ecore_window_property_get(win, a_win_layer, XA_CARDINAL, &size);
|
2000-12-08 14:54:42 -08:00
|
|
|
if (props)
|
|
|
|
|
{
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
int num;
|
2000-12-08 14:54:42 -08:00
|
|
|
|
|
|
|
|
num = size / sizeof(int);
|
|
|
|
|
if (num > 0) b->client.layer = props[0];
|
|
|
|
|
FREE(props);
|
|
|
|
|
}
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
2001-01-02 15:10:12 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_get_title(Window win, E_Border *b)
|
|
|
|
|
{
|
|
|
|
|
char *title;
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
title = ecore_window_get_title(win);
|
2001-01-02 15:10:12 -08:00
|
|
|
|
|
|
|
|
if (b->client.title)
|
|
|
|
|
{
|
2001-10-18 19:02:56 -07:00
|
|
|
if ((title) && (!strcmp(title, b->client.title)))
|
|
|
|
|
{
|
|
|
|
|
FREE(title);
|
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;
|
2001-10-18 19:02:56 -07:00
|
|
|
}
|
2001-01-02 15:10:12 -08:00
|
|
|
b->changed = 1;
|
|
|
|
|
FREE(b->client.title);
|
|
|
|
|
}
|
|
|
|
|
b->client.title = NULL;
|
|
|
|
|
if (title) b->client.title = title;
|
2001-08-16 01:45:37 -07:00
|
|
|
else e_strdup(b->client.title, "No Title");
|
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;
|
2001-01-02 15:10:12 -08:00
|
|
|
}
|
|
|
|
|
|
2001-10-07 23:53:26 -07:00
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_class(Window win, E_Border *b)
|
|
|
|
|
{
|
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;
|
|
|
|
|
|
2001-10-07 23:53:26 -07:00
|
|
|
IF_FREE(b->client.name);
|
|
|
|
|
IF_FREE(b->client.class);
|
|
|
|
|
b->client.name = NULL;
|
|
|
|
|
b->client.class = NULL;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_get_name_class(win, &(b->client.name), &(b->client.class));
|
2001-10-07 23:53:26 -07:00
|
|
|
if (!b->client.name) e_strdup(b->client.name, "Unknown");
|
|
|
|
|
if (!b->client.class) e_strdup(b->client.class, "Unknown");
|
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;
|
2001-10-07 23:53:26 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_hints(Window win, E_Border *b)
|
|
|
|
|
{
|
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;
|
|
|
|
|
|
2001-11-15 21:39:34 -08:00
|
|
|
ecore_window_get_hints(win,
|
|
|
|
|
&(b->client.takes_focus),
|
|
|
|
|
&(b->client.initial_state),
|
|
|
|
|
NULL, NULL, NULL,
|
|
|
|
|
&(b->client.group));
|
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;
|
2001-10-07 23:53:26 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_machine(Window win, E_Border *b)
|
|
|
|
|
{
|
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;
|
|
|
|
|
|
2001-10-07 23:53:26 -07:00
|
|
|
IF_FREE(b->client.machine);
|
|
|
|
|
b->client.machine = NULL;
|
2001-10-17 15:34:27 -07:00
|
|
|
b->client.machine = ecore_window_get_machine(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;
|
2001-10-07 23:53:26 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_command(Window win, E_Border *b)
|
|
|
|
|
{
|
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;
|
|
|
|
|
|
2001-10-07 23:53:26 -07:00
|
|
|
IF_FREE(b->client.command);
|
|
|
|
|
b->client.command = NULL;
|
2001-10-17 15:34:27 -07:00
|
|
|
b->client.command = ecore_window_get_command(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;
|
2001-10-07 23:53:26 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_icon_name(Window win, E_Border *b)
|
|
|
|
|
{
|
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;
|
|
|
|
|
|
2001-10-07 23:53:26 -07:00
|
|
|
IF_FREE(b->client.icon_name);
|
|
|
|
|
b->client.icon_name = NULL;
|
2001-10-17 15:34:27 -07:00
|
|
|
b->client.icon_name = ecore_window_get_icon_name(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;
|
2001-10-07 23:53:26 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_get_state(Window win, E_Border *b)
|
|
|
|
|
{
|
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_RETURN;
|
2001-11-15 21:39:34 -08:00
|
|
|
UN(win);
|
|
|
|
|
UN(b);
|
2001-10-07 23:53:26 -07:00
|
|
|
}
|
|
|
|
|
|
2000-12-08 14:54:42 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_set_frame_size(Window win, int l, int r, int t, int b)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_e_frame_size = 0;
|
|
|
|
|
int props[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_ENTER;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_e_frame_size, "_E_FRAME_SIZE");
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = l;
|
|
|
|
|
props[1] = r;
|
|
|
|
|
props[2] = t;
|
|
|
|
|
props[3] = b;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_e_frame_size, XA_CARDINAL, 32, props, 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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_set_desk_area(Window win, int ax, int ay)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_win_area = 0;
|
|
|
|
|
int props[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_area, "_WIN_AREA");
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = ax;
|
|
|
|
|
props[1] = ay;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_win_area, XA_CARDINAL, 32, props, 2);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_set_desk_area_size(Window win, int ax, int ay)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_win_area_count = 0;
|
|
|
|
|
int props[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_area_count, "_WIN_AREA_COUNT");
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = ax;
|
|
|
|
|
props[1] = ay;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_win_area_count, XA_CARDINAL, 32, props, 2);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_set_desk(Window win, int d)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_win_workspace = 0;
|
|
|
|
|
int props[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_workspace, "_WIN_WORKSPACE");
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = d;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_win_workspace, XA_CARDINAL, 32, props, 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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
2000-12-18 13:28:44 -08:00
|
|
|
int
|
|
|
|
|
e_icccm_is_shaped(Window win)
|
|
|
|
|
{
|
|
|
|
|
int w, h, num;
|
|
|
|
|
int shaped = 1;
|
|
|
|
|
XRectangle *rect;
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_get_geometry(win, NULL, NULL, &w, &h);
|
|
|
|
|
rect = ecore_window_get_shape_rectangles(win, &num);
|
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 (!rect)
|
|
|
|
|
D_RETURN_(1);
|
|
|
|
|
|
2000-12-18 13:28:44 -08:00
|
|
|
if ((num == 1) &&
|
|
|
|
|
(rect[0].x == 0) && (rect[0].y == 0) &&
|
|
|
|
|
(rect[0].width == w) && (rect[0].height == h))
|
|
|
|
|
shaped = 0;
|
|
|
|
|
XFree(rect);
|
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_(shaped);
|
2000-12-18 13:28:44 -08:00
|
|
|
}
|
|
|
|
|
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_handle_property_change(Atom a, E_Border *b)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_wm_normal_hints = 0;
|
2001-01-02 15:10:12 -08:00
|
|
|
static Atom a_motif_wm_hints = 0;
|
|
|
|
|
static Atom a_wm_name = 0;
|
2001-10-07 23:53:26 -07:00
|
|
|
static Atom a_wm_class = 0;
|
|
|
|
|
static Atom a_wm_hints = 0;
|
|
|
|
|
static Atom a_wm_client_machine = 0;
|
|
|
|
|
static Atom a_wm_command = 0;
|
|
|
|
|
static Atom a_wm_icon_name = 0;
|
|
|
|
|
static Atom a_wm_state = 0;
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_wm_normal_hints, "WM_NORMAL_HINTS");
|
|
|
|
|
ECORE_ATOM(a_motif_wm_hints, "_MOTIF_WM_HINTS");
|
|
|
|
|
ECORE_ATOM(a_wm_name, "WM_NAME");
|
|
|
|
|
ECORE_ATOM(a_wm_class, "WM_CLASS");
|
|
|
|
|
ECORE_ATOM(a_wm_hints, "WM_HINTS");
|
|
|
|
|
ECORE_ATOM(a_wm_client_machine, "WM_CLIENT_MACHINE");
|
|
|
|
|
ECORE_ATOM(a_wm_command, "WM_COMMAND");
|
|
|
|
|
ECORE_ATOM(a_wm_icon_name, "WM_ICON_NAME");
|
|
|
|
|
ECORE_ATOM(a_wm_state, "WM_STATE");
|
2001-10-07 23:53:26 -07:00
|
|
|
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
if (a == a_wm_normal_hints) e_icccm_get_size_info(b->win.client, b);
|
2001-01-02 15:10:12 -08:00
|
|
|
else if (a == a_motif_wm_hints) e_icccm_get_mwm_hints(b->win.client, b);
|
|
|
|
|
else if (a == a_wm_name) e_icccm_get_title(b->win.client, b);
|
2001-10-07 23:53:26 -07:00
|
|
|
else if (a == a_wm_class) e_icccm_get_class(b->win.client, b);
|
|
|
|
|
else if (a == a_wm_hints) e_icccm_get_hints(b->win.client, b);
|
|
|
|
|
else if (a == a_wm_client_machine) e_icccm_get_machine(b->win.client, b);
|
|
|
|
|
else if (a == a_wm_command) e_icccm_get_command(b->win.client, b);
|
|
|
|
|
else if (a == a_wm_icon_name) e_icccm_get_icon_name(b->win.client, b);
|
|
|
|
|
else if (a == a_wm_state) e_icccm_get_state(b->win.client, b);
|
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;
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
2001-10-17 15:34:27 -07:00
|
|
|
e_icccm_handle_client_message(Ecore_Event_Message *e)
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -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_RETURN;
|
Sorry guys.. I had to revert a bunch of changes.. that's life.. but READ the
following (it's in the README now)
-------------------------------------------------------------------------------
Enlightenment 0.17.0 CVS Code....
-------------------------------------------------------------------------------
The Rasterman - raster@valinux.com, raster@rasterman.com
*******************************************************************************
**************** READ THIS! It is of the UTMOST IMPORTANCE! *******************
*******************************************************************************
This is the source code for Enlightenment 0.17 - If you got this you got it
from Enlightenment's CVS repository - or from someone who took it out of
the CVS repository.
The CVS repository is full of code *IN DEVELOPMENT* - that often means it's
in the middle of being worked on and may install strange things in strange
places, make a mess, and may not even be compatible with a final release. If
you at all use this code, you are HEAVILY URGED, when it is finally released,
to remove all traces of anything this CVS code base has installed on your
system (it is COMPLETELY up to you to keep track of that - do NOT expect any
help), and then install the full release on a cleaned system. Don't come
asking "can I just keep using CVS" oonce things are released - thqat is the
reason I pu this paragraph here - so you don't ask. The asnwer is the same
as above - if there is a proper final release use that. CVS is really only
for those havily hacking on the code.
Now we have that warning over and done with. How to build and install from
CVS?
$ ./autogen.sh && make
$ su
Password:
<- as root ->
# make install
You should be able to use the binary of enlightenment as a window manager.
you might be advised for cleanliness to do
$ ./autogen.sh --prefix=/usr/local/e-17
so it installs relative to the /usr/local/e-17 directory and keeps all the
e-17 development code and data in that tree so it is easily removed when the
time codes.
NOTES: Read these carefully!
Enlightenment does not check for previously running Window Managers right
now - so you need to make sure no other WM is running - E will not do that
for you.
Enlightenment has no menus or keybindings or any way of launching
applications right now - you'll have to figure out an alternative way of
doing it.
Enlightenment only handles a small subset of ICCCM and thus will have bugs -
some applications will not behave correctly and may apear in odd spots or
not resize or place themselves properly etc. Expect this - it's code being
worked on. Just be happy it does as much as it already does.
Enlightenment RELIES on lots of libraires that have been written. Ecore,
Ebits, Evas, Edb, Imlib2 just to mention a few. Especially Ebits, Ecore and
Evas change in CVS often - you will need the absolute latest of these if you
wish Enlightenment 0.17 code to run properly or compile. If you update
Enlightenment from CVS update these too to get any changes they have in
their trees.
If you plan on working on the code... STOP! don't rush in and work on it -
even if you have CVS commit access - EXPECT me (Raster) to revert any changes
you make if you do this - regardless of the changes and how much work you
put into them. First read the code well and LEARN it. If you have questions
about some of the more obscure hidden program flow - ASK - but don't go
tampering with it - Enlightenment 0.17's code is much more complex and
intricate than E 0.16 - but at the same time it's much cleaner and more
object oriented. Learn it well first. Some parts of E 0.17 are "hacked" with
hard-coded stuff, just so, for now, it works. They will be virtualized and
imporved over time. If you have plans - tell me about them first - discuss
them before you go impliment them. I know I already have a lot of the
components of E 0.17's code planned in my head - but I won't get to them for
a while - and if people go impliment or hack bad stuff in, it means I have to
spend lots of time fixing something that is bad in the first place, or we
end up doing duplicate work. There *IS* a plan - believe it or not - but to
be honest - it's more complex and large than I can just write down in a
README, so talk about your ideas first. I'm going to be ruthless in keeping
the code neat, clean and free of nasty hacks (except ones I put in as
temporary stop-gap measures to make the thing work - since I know where
those are and what I need to do to do it right). If you can't find me or I
don't reply to your e-mail - don't get impatient - just wait. I currently
have no network access at home, so I'm doing a chunk of code offline. I'll
get to your mail and queries as time allows.
If you have problems with the code or bugs to report, kindly forward them to
/dev/null (the code is in now way or form ready for bug reports - I don't
want crap filling my mailbox).
I hope that clears things up for now.
SVN revision: 3976
2000-12-11 12:08:38 -08:00
|
|
|
UN(e);
|
|
|
|
|
}
|
|
|
|
|
|
2000-12-08 14:54:42 -08:00
|
|
|
void
|
|
|
|
|
e_icccm_advertise_e_compat(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;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
D_RETURN;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_advertise_mwm_compat(void)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_motif_wm_info = 0;
|
|
|
|
|
int props[2];
|
|
|
|
|
|
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;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_motif_wm_info, "_MOTIF_WM_INFO");
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = 2;
|
2001-10-17 15:34:27 -07:00
|
|
|
props[0] = ecore_window_root();
|
|
|
|
|
ecore_window_property_set(0, a_motif_wm_info, a_motif_wm_info, 32, props, 2);
|
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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_advertise_gnome_compat(void)
|
|
|
|
|
{
|
|
|
|
|
static Atom a_win_supporting_wm_check = 0;
|
|
|
|
|
static Atom a_win_protocols = 0;
|
|
|
|
|
static Atom a_win_wm_name = 0;
|
|
|
|
|
static Atom a_win_wm_version = 0;
|
|
|
|
|
static Atom a_win_layer = 0;
|
|
|
|
|
int props[32];
|
|
|
|
|
Window 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_ENTER;
|
|
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_protocols, "_WIN_PROTOCOLS");
|
|
|
|
|
ECORE_ATOM(a_win_layer, "_WIN_LAYER");
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = a_win_protocols;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(0, a_win_protocols, XA_ATOM, 32, props, 1);
|
2000-12-08 14:54:42 -08:00
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_wm_name, "_WIN_WM_NAME");
|
|
|
|
|
ecore_window_property_set(0, a_win_wm_name, XA_STRING, 8, "Enlightenment", strlen("Enlightenment"));
|
|
|
|
|
ECORE_ATOM(a_win_wm_version, "_WIN_WM_VERSION");
|
|
|
|
|
ecore_window_property_set(0, a_win_wm_version, XA_STRING, 8, "0.17.0", strlen("0.17.0"));
|
2000-12-08 14:54:42 -08:00
|
|
|
|
2001-10-17 15:34:27 -07:00
|
|
|
ECORE_ATOM(a_win_supporting_wm_check, "_WIN_SUPPORTING_WM_CHECK");
|
|
|
|
|
win = ecore_window_override_new(0, 0, 0, 7, 7);
|
2000-12-08 14:54:42 -08:00
|
|
|
props[0] = win;
|
2001-10-17 15:34:27 -07:00
|
|
|
ecore_window_property_set(win, a_win_supporting_wm_check, XA_CARDINAL, 32, props, 1);
|
|
|
|
|
ecore_window_property_set(0, a_win_supporting_wm_check, XA_CARDINAL, 32, props, 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;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_advertise_kde_compat(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;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
D_RETURN;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
e_icccm_advertise_net_compat(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;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
D_RETURN;
|
2000-12-08 14:54:42 -08:00
|
|
|
}
|