otherwise we are leaking a eina list.
This was discovered while running the elm_suite with CK_FORK=no, since
ecore and eina are init´ed and shutdown´ed, after the shutdown this list
points to freeed memory pools. So in the next testcase the list is
invalid and crashes. This fixes it with freeing the list on shutdown.
Extend edje edit deberate source API. Add two new API.
- edje_edit_object_color_class_list_get return a list of used color classes for
given object
- edje_edit_color_classes_source_generate is generate code for given
color classes list
This API is annex for edje_edit_object_source_generate. Together this
API's provide a mechanism for generate source for custom groups from
a edj file or more.
Lacking a proper internal tag, I'm using both protected (it is
in fact a protected access function) and beta (to mark as unstable,
not real API).
New smart objects based on EO only should rely on constructor,
finalize and destructor exclusively. In theory, this should be fine.
Unfortunately it may be impossible to inherit from the Efl.Ui.Win
class as it uses a really bad hack and calls super.constructor
inside the finalize method.
This is an override of efl_gfx_size_set. Same as before, the
order of operations matter so it is possible that a corner
case will break. In particular, legacy code was:
- intercept
- smart resize (do stuff), super, super, super
- evas object resize
The new code is more like:
- intercept
- super, super, super, evas object resize
- do stuff
But unfortunately this broke elm_widget (read: all widgets) as
the internal resize was done before the object resize. So,
inside the resize event cb, the resize_obj size would not match
the smart object size. >_<
This is an override of efl_gfx_position_set.
As for the other patches, I hope I didn't break anything.
A problem likely to happen is that the super call was inserted
too early or too late in the call flow. For instance:
_myclass_position_set(obj, x, y) {
position_set(super(obj), x, y);
position_get(obj, &prevx, &prevy);
do_something_with_delta_xy();
}
The above code flow is obvisouly wrong, but may have crept in this
patch (such a bug sneaked in inside smart object, breaking
everything at first).
While this kind of API seems to make sense with smart objects
(relative coordinates), it is currently not used apart from
the smart object class itself.
So, for now, I'm moving this to legacy to clean up Efl.Canvas.Group
and we can later add the equivalent in a clean "group" API.
These should be just overrides of Efl.Gfx.visible.set. Many
widgets were handling smart show() and hide() manually, which
means this patch is quite large.
Hopefully this doesn't break anything, obviously. But here are
some widgets known to be problematic, as the old code flow was
really strange (sometimes not calling the efl_super function):
- window
- notify
Similarly to group_color_set, group_clip_[un]set should not
exist and should be a result of efl_super and inheritance.
This patch also removes clip_unset from the EO API and keeps
only clip_set(NULL). The reason is that it will avoid bad overrides
of clip_unset() vs. clip_unset(NULL). This also simplifies the code
a bit. Ideally we should be able to reintroduce clip_unset in EO
if we can have a "@final" tag (like java's final keyword), to
prevent overrides.
Widgets should simply override efl_gfx_color_set and call
super all the way up to evas object.
Legacy compatibility with call interceptors and early call
abortion (eg. delete_me or obj->layer == NULL) are implemented
with an internal call. See the previous commit introducing the
API.
This is a poor man's solution to get rid of group functions such
as clip_set, clip_unset, color_set, etc... See the following
commits.
This API needs to be EAPI for elementary but shouldn't be used
outside EFL. This is required purely for legacy compatibility.
Here's the call flow, inside show(obj):
1. if (intercept_show(obj)) return;
2. show(super(obj));
3. do other stuff
Most of the smart functions "Efl.Canvas.Group.group_xxx" should
not exist and be overrides of the base object function instead.
Cleaning this up is necessary if we want an EO API for custom
smart objects. This patch is the first attempt at removing a
method (the simplest one).
As for no_render, I wonder if propagating to the children
really is necessary. evas_render should skip them already.
When we destroy outputs, we should be freeing the Output's Modes also
as that was previously allocated memory.
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary: when transition animation is working, sometimes one pixels loss occurs.
Reviewers: cedric, woohyun, jpeg, raster
Subscribers: jpeg
Differential Revision: https://phab.enlightenment.org/D4341
Summary:
If the pixels of image object has updates
via native_suface_set api with TBM surface,
does not update on screen(no rendering)
Reviewers: jpeg, cedric
Differential Revision: https://phab.enlightenment.org/D4340
"item,activated" smart callback is added to notice when an item is
activated.
"item,activated" smart callback is called as follows.
1. Item push is finished and the new item is activated.
2. Item pop is finished and the previous item is activated.
void *event_info in smart callback is the activated item pointer.
This happens with shared objects.
The situation seems to be:
1. object has composited object a of class A in thread 1
2. call something on object a from thread 2, deadlock
In fact, do anything from thread 2 on a shared object and you deadlock.
EO is now extremely restrictive wrt. threads so that efl_data_scope_get()
can't work outside the main loop. This patch fixes the usage to create
sw buffers as shared objects (accessible from both the main loop and evas
async thread) and use plain old pointers where possible.
The buffers now have no parent because efl_add(CLASS, obj_from_mainloop)
does not work with shared objects. This is bad, as the buffers conceptually
belong to the main loop, and only need to be accessible from the draw thread
for a few calls. The main loop determines their lifecycle.
Fixes T4628
this fixes intitial iconic state for x11 as demonstrated by
terminology -I
but enlightenment is broken though... xterm -iconic also shows the
same break with a black window.
@fix
so we handled override cases and set withdrawn to false on show, but
when normally managed it might be nicer to wait for a state change via
the wm state property to know we are "normal"
this should fix T4699
@fix
The value stored by efl_future_all is an accessor, not an
iterator. This fixes the usage of promises internally to
the fileselector widget.
See 1a2014a122
Fixes T4686
An eio model (and I believe most other models as well, in fact)
require a Efl.Loop.User as parent in order for efl future/promises
to work. This adds a fallback code directly inside eio model. The
alternative was to parent and efl_del all models in fileselector.
See T4686
And indirectly also Efl.Canvas.Object.
I believe those two classes should even inherit from Efl.Loop.User.
Right now this patch relies on the new dependence of Evas over Ecore,
and is maybe a bit ugly as is.
Ping @cedric
See T4686
This reverts commit 2c736adc87.
well that was totally unexpected. - efl app windows dont show at all..
wtf? this should not have affected that at all..
so we handled override cases and set withdrawn to false on show, but
when normally managed it might be nicer to wait for a state change via
the wm state property to know we are "normal"
this should fix T4699
We need a method that allows us to place the cursors at the start and end of an
annotation. This is required for things like getting the geometry of a range.
@feature
The ported geometry_get was actually the legacy simple_geometry_get.
For getting simple geometries like selection this was enough, but I forgot that
we also need to query more complex geometries e.g. links.
This is required to implement link anchors in Ui Text.
Now geometry_get and simple_geometry_get are the same as their legacy
counterparts.
@feature
It's not because the bug with __builtin_prefetch is inside
clang/llvm that we must break the build for people who prefer it
over gcc. As soon as a non-broken version is out, the ifdef must
be either removed (and ask people to update their clang install)
or add a version check based on __clang_xxx__.
Compilation tested with clang 3.8.1 and gcc 6.2.1.
"push,finished" and "pop,finished" smart callbacks are added to notice
when push and pop operation is finished.
New smart callback signals are added as follows.
"push,finished,item,activated" : Called when item push is finished and
void *event_info in smart callback is activated item (new item).
"push,finished,item,deactivated" : Called when item push is finished and
void *event_info in smart callback is deactivated item (current item).
"pop,finished,item,activated" : Called when item pop is finished and
void *event_info in smart callback is activated item (previous item).
"pop,finished,item,deactivated" : Called when item pop is finished and
void *event_info in smart callback is deactivated item (current item).
so ecore_x_vsync as a tool uses glx for nvidia drivers plut "wait for
vblank" extensions to try vsync "sync". the problem is this is flakey
because the drivers may or may not continue vsyncing after screen off
or syspend/resume or vt changes and all the workarounds dont seem to
be reliable, so since this causes this to be disabled, no point
keeping all the code and build stuff around, so remove this "unused
junk" we have in the tree.
macOS Sierra provides clock_gettime().
Instead of testing for the Mach kernel to re-implement clock_gettime(),
let first check that clock_gettime() is not already declared. Older OSX
versions will have the Mach-only fallback.
@fix
i see a speedup of about 8% over a series of list walking and freeing
functions given this change. it's a small speedup but still not too
shabby just for some prefetches thrown in. ymmv depending on memory
subsystem, memory speed itself, cpu and architecture.
@optimize
this allows you to portably use prefetch compiler builtins. this adds
EINA_PREFETCH(), EINA_PREFETCH_WRITE(), EINA_PREFETCH_NOCACHE() and
EINA_PREFETCH_NOCACHE_WRITE() macros to do this that are "nothing" if
your compiler doesnt support it. of course it also requires your
compielr compile instructions for your architecture and it can only do
so if the architecture it compiles for has these instructions, so be
aware.
@feat
Summary:
It should be evas_device_add_full() in order to follow the EFL
name pattern.
Reviewers: DaveMDS, bdilly
Reviewed By: DaveMDS, bdilly
Subscribers: cedric, jpeg
Differential Revision: https://phab.enlightenment.org/D4325
this moves a lot of error case handling into goto's so the code gets
out of the hot path and this should help expecially since variou
smacros do things like:
do { char buf[256]; sprintf(buf, fmt, ptr); _eo_pointer_error(buf); } while (0)
_Efl_Class *klass; \
do { \
klass = _eo_class_pointer_get(klass_id); \
if (!klass) { \
_EO_POINTER_ERR("Class (%p) is an invalid ref.", klass_id); \
return ret; \
} \
} while (0)
so putting quite a chunk of code inside a rare "if this errors"
handler that will cause l1 cache misses and this we don't want, thus
moving stuff in eo core out of hot paths to cut down on overhead. yes
it might not be pretty but it's kind of the right thing at such a core
level of efl. this also does the same to the eo base class as this is
also going to be relatively hot given it's the core of every other
object.
In order to avoid the following build warning:
lib/elementary/elm_atspi_bridge.c: In function
‘_children_changed_signal_send’:
lib/elementary/elm_atspi_bridge.c:3971:4: warning: ‘idx’ may
be used uninitialized in this function [-Wmaybe-uninitialized]
_bridge_signal_send(data, event->object, ATSPI_DBUS_INTERFACE_EVENT_OBJECT
A little refactoring was done to function _children_changed_signal_send.
Actually I wasn't able to see a codepath to use idx unintialized,
but this way is more clear and avoid compiler possible confusion.
Remove codegen_example_generated.h from codegen_example_SOURCES
and let it only on nodist_codegen_example_SOURCES and
on BUILT_SOURCES.
Also add dependency between codegen_example.c
and codegen_example_generated.h since it's required
to compile.
Avoid the following build error:
CODEGEN codegen_example_generated.c
codegen_example.c:26:39: fatal error: codegen_example_generated.h:
No such file or directory
compilation terminated.
Makefile:4960: recipe for target 'codegen_example.o' failed
Fix missing dependency.
Get rid of the following error:
/usr/bin/ld: ecore_evas_vnc.o: undefined reference
to symbol 'ECORE_EVENT_MOUSE_BUTTON_UP'
src/lib/ecore_input/.libs/libecore_input.so.1: error adding symbols:
DSO missing from command line
collect2: error: ld returned 1 exit status
Makefile:2306: recipe for target 'ecore_evas_vnc' failed
Summary:
The inherit_only groups could be used for template of other groups.
So, even if it has image part, it dose not need to have image attributes.
@fix
Test Plan: N/A
Reviewers: cedric, conr2d, jpeg, raster
Reviewed By: raster
Differential Revision: https://phab.enlightenment.org/D4322
As ector objects are acessed by draw thread we need to create it as
shared object in order to access it from other thread.
Note: there is some performance lag...
Summary: make ector object as shared eo object to acess from other thread.
Reviewers: cedric, jpeg, raster
Reviewed By: jpeg, raster
Subscribers: cedric, jpeg
Differential Revision: https://phab.enlightenment.org/D4319
so there were a few issues. one we had a spinlokc on the eoid table
for shared objects AND then had a mutex for accessing those objects
(released on return from any eo function). BUT this missed some funcs
like eo_ref, eo_unref and so on in eo.c ... oops. so fixed. but then i
realized there was a race condition. we locked the eoid table then
unlocked with our pointer THEN locked the sharted object mutex ...
then unlocked it. that was a race condtion gap. so we should share the
same lock anyway - if it's a shared object, grab the shared object
mutex then do a lookup and if the lookup does not fail, KEEP the lock
until it is released by the return from eo function or by some special
macro/funcs that released a matching lock. since its a recursive lock
this is all fine. as its also a universal single lock for all objects
we just need the eoid to know if it's shared and needs locking based
on the domain bits. so now do this locking properly with just a single
mutex, not both a spinlock and mutex and keep the lock around until
totally done with the object. this plugs the race condition holes and
goes from 1 spinlock lock and unlock then a mutex lock and unlokc to
just a single mutex lock and unlock. this means shared objects are
actually truly safe across threads and only have the overhead of a
single recursive mutex to lock and unlock in every api call.
Summary: Need to NULL check before using surface like other cases.
Test Plan: N/A
Reviewers: jpeg, raster, ManMower, devilhorns
Reviewed By: devilhorns
Subscribers: cedric
Differential Revision: https://phab.enlightenment.org/D4301
Summary:
1. Fix wrong logic of ELM_PANEL_ORIENT_RIGHT case in _state_sync.
2. Move pre calc function in anim_cb to sizing_eval function.
This code have caused problem that scrollable panel is not animated during drawer open and close.
3. Maintain content size ratio when orient is changed.
4. Freeze scroller when orient is changed.
This code will need to change orientation when drawer opened.
Reviewers: cedric, eunue
Subscribers: cedric, jpeg
Differential Revision: https://phab.enlightenment.org/D4320
this is quick to get evas working again. this disables part of
7e2d700d06 that creates deadlocks. this
thread waits on mainloop. mainloop waits on this render thread. BOOM.
deadlock.
This patch adds the support for Ecore events from a remove
VNC client. Every time it happens a VNC mouse move/click/wheel or a
VNC keyboard event an Ecore event event will be created and dispatched.
The Efl_Input_Device will contain the device which originated the event,
this will be useful for dealing with multi-seat environment. When the device
is NULL it means that the event originated from the default seat.
Summary:
Active descendant changed signal send function checks for wrong object,
hence always returns with out actually sending signal.
The state is set to pd->object_broadcast_mask and not
pd->object_children_broadcast_mask in _set_broadcast_flag function.
Test Plan:
Send the active descandant changed signal from any widget, but signal
is not sent.
Reviewers: kimcinoo, lukasz.stanislawski, cedric
Reviewed By: cedric
Subscribers: jpeg, govi, rajeshps
Differential Revision: https://phab.enlightenment.org/D4318
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
If we fail to allocate memory for savedtypes.types here, then the
"known" array would end up leaking. Defer creation of 'known' array
until after savedtypes.types is allocated, this way we don't leak.
Fixes CID1363216
Signed-off-by: Chris Michael <cp.michael@samsung.com>
Summary:
If entry edje should be changed according to scrollable mode,
the signal will be used. The following signals are added.
"elm,scroll,enable"
"elm,scroll,disable"
Test Plan: N/A
Reviewers: raster, tasn, herdsman, cedric
Subscribers: minkyu, jpeg, akanad, z-wony, Blackmole
Differential Revision: https://phab.enlightenment.org/D4254
if we dont get the userdata of a offer something in the protocol went
very very bad, we should not emit there a error, just return.
Also if we are getting a null offer we should dereference the nulled out
offer object.
This fixes CID 1363214
Since _config_load() is called and _elm_config is not initialized, NULL
check for _elm_config is not necessary.
The unnecessary NULL check causes false alarm because _elm_config is not
checked in other cases.
so by chance i discovered efreet is doing bad things(tm) when
expanding/appending to string buffers to generate commands based off
desktop files. the string append basically was buggy, so fixed it by
making it a lot simpler and more obvious and now reliable.
@fix
The callbacks are never invoked due to sd->play equal to 0. The
function _emotion_decode_stop is called before and resets this field.
Before the change to Efl.Canvas.Video, sd->play was not checked.
if a dnd operation enters a surface the window_id is clear, so a offer
is always specific to a window. If we have a source we try to fetch the
id from the focus/prevfocus or keyboardfocus.
This fixes dropping into a efl app which never got focus before.
the format can also be the result of mutliple or´ed values.
The new code now also uses the same mimetypes to format type relation
than the selection code.
This fixes dragging onto a container with multiple formats.
Spotted by ApBBB while dragging from ephoto to terminology.
This also fixes T3320
as per other recent benchmarking, moving rearely run code (in this
case code to init the op etc.) out of the l1 cacheline prefetch inot a
blob of code at the end of the function where we goto and goto back
again should provide decent-ish speedups for the resolv cache in
avoding this code. yes it makes the code less pretty to read but at
this really low level hot path ... evil things must happen to get the
speed we want/need.
As we always set this flag in the drm2_fb_flip function, having this
check here is now pointless.
Signed-off-by: Chris Michael <cp.michael@samsung.com>