platform/upstream/efl.git
11 years agoevas/cserve2: use Eina_Stringshare.
Jean-Philippe Andre [Mon, 1 Jul 2013 08:41:13 +0000 (17:41 +0900)]
evas/cserve2: use Eina_Stringshare.

We must close the image after loading the data.
Also, Evas loaders expect stringshares.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas: fix compilation warning in evas_font_draw.
Jean-Philippe Andre [Tue, 2 Jul 2013 06:18:19 +0000 (15:18 +0900)]
evas: fix compilation warning in evas_font_draw.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: return in case of select timeout.
Jean-Philippe Andre [Mon, 1 Jul 2013 01:55:46 +0000 (10:55 +0900)]
evas/cserve2: return in case of select timeout.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: add some more error debug (client side).
Jean-Philippe Andre [Tue, 2 Jul 2013 05:29:47 +0000 (14:29 +0900)]
evas/cserve2: add some more error debug (client side).

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/loaders: init/shutdown EET from loader module.
Jean-Philippe Andre [Fri, 28 Jun 2013 08:02:07 +0000 (17:02 +0900)]
evas/loaders: init/shutdown EET from loader module.

cserve2 slave requires Eet to be initialized before use :)

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/loaders: add log domain for ICO loader.
Jean-Philippe Andre [Thu, 27 Jun 2013 07:44:58 +0000 (16:44 +0900)]
evas/loaders: add log domain for ICO loader.

This is to remove the dependency of the ico loader on Evas itself.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: remove dead code (old loaders API).
Jean-Philippe Andre [Mon, 1 Jul 2013 08:52:06 +0000 (17:52 +0900)]
evas/cserve2: remove dead code (old loaders API).

Remove: cserve-specific loaders and deprecated code.
Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: fix crash with YUV tests in expedite.
Jean-Philippe Andre [Mon, 1 Jul 2013 08:03:31 +0000 (17:03 +0900)]
evas/cserve2: fix crash with YUV tests in expedite.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: use Evas loaders from slave
Jean-Philippe Andre [Thu, 27 Jun 2013 07:18:40 +0000 (16:18 +0900)]
evas/cserve2: use Evas loaders from slave

Use common loaders with evas, and ignore specific modules.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/modules: Add preprocessor ifs
Jean-Philippe Andre [Tue, 2 Jul 2013 02:00:44 +0000 (11:00 +0900)]
evas/modules: Add preprocessor ifs

We'll want to distinguish full builds for evas and partial
builds for cserve2 slaves.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas: build evas_module within cserve slave
Jean-Philippe Andre [Thu, 27 Jun 2013 06:31:20 +0000 (15:31 +0900)]
evas: build evas_module within cserve slave

Objective: use common loaders from cserve2

Ideally evas_module should be a static library but it would
then require static/dynamic linking to too many modules unused
by cserve2 (eg. engines & savers)

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: avoid useless SETOPT messages.
Jean-Philippe Andre [Wed, 26 Jun 2013 08:47:41 +0000 (17:47 +0900)]
evas/cserve2: avoid useless SETOPT messages.

Most of the time, the SETOPT message is sent only
to carry 0 data. Let's add a flag to the OPEN message to
avoid useless socket calls.
Server-side: implement with message faking.
Also, fix image ID logic on client side:
 - alloc data entry asap,
 - assign image_id during call to OPEN

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: Simplify debug logs (server side).
Jean-Philippe Andre [Wed, 26 Jun 2013 08:19:49 +0000 (17:19 +0900)]
evas/cserve2: Simplify debug logs (server side).

Remove newlines.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: Set default LRU size to 4Mb or env-based
Jean-Philippe Andre [Wed, 26 Jun 2013 07:31:38 +0000 (16:31 +0900)]
evas/cserve2: Set default LRU size to 4Mb or env-based

The LRU should not be limited to 0 byte otherwise all scaled
images will be dropped and reloaded constantly, killing the
performance.
The size is inspired from the scalecache's size of 4Mb.

Variable: EVAS_CSERVE2_SIZE (number in Mb)
Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: pass loader data during open.
Jean-Philippe Andre [Wed, 26 Jun 2013 06:11:06 +0000 (15:11 +0900)]
evas/cserve2: pass loader data during open.

Note: this path seems never to be used.
Loader options are set by the file loader or after open.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: write() can return other errors.
Jean-Philippe Andre [Wed, 26 Jun 2013 02:50:07 +0000 (11:50 +0900)]
evas/cserve2: write() can return other errors.

In case of error, return from the function and don't decrement
the send variable.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: minor fixes (error paths).
Jean-Philippe Andre [Wed, 26 Jun 2013 02:23:35 +0000 (11:23 +0900)]
evas/cserve2: minor fixes (error paths).

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas/cserve2: use null-terminated strings everywhere.
Jean-Philippe Andre [Wed, 19 Jun 2013 06:13:25 +0000 (15:13 +0900)]
evas/cserve2: use null-terminated strings everywhere.

Some hashtable lookups failed because the keys were added
either with eina_stringshare_add() or e_s_add_lenth() but
looked for with e_s_add_length() only.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas - updated missed CNN
ChunEon Park [Tue, 2 Jul 2013 06:47:13 +0000 (15:47 +0900)]
evas - updated missed CNN

11 years agoevas/proxy rendering - optimize proxy rendering.
ChunEon Park [Tue, 2 Jul 2013 06:38:42 +0000 (15:38 +0900)]
evas/proxy rendering - optimize proxy rendering.

re-render proxies if sources are changed. this patch increases the proxy rendering up to 2x.

11 years agoevas/render - refactoring about checking smart member changed.
ChunEon Park [Tue, 2 Jul 2013 06:28:56 +0000 (15:28 +0900)]
evas/render - refactoring about checking smart member changed.

move those code to the evas_object_smart.

11 years agoefl: cleanup NEWS.
Cedric Bail [Tue, 2 Jul 2013 05:50:02 +0000 (14:50 +0900)]
efl: cleanup NEWS.

11 years agoeet: force thread to always run during eet_cache_concurrency test.
Cedric Bail [Tue, 2 Jul 2013 04:45:30 +0000 (13:45 +0900)]
eet: force thread to always run during eet_cache_concurrency test.

11 years agoedje_text: Clarify that the block and else are independent
Daniel Willmann [Mon, 1 Jul 2013 15:04:27 +0000 (16:04 +0100)]
edje_text: Clarify that the block and else are independent

Sometimes whitespace helps

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoFixed ChaneLog formatting.
Tom Hacohen [Mon, 1 Jul 2013 13:19:05 +0000 (14:19 +0100)]
Fixed ChaneLog formatting.

11 years agoEvas textblock: Fixed format parsing to not be confused by single-quotes.
Tom Hacohen [Mon, 1 Jul 2013 13:11:27 +0000 (14:11 +0100)]
Evas textblock: Fixed format parsing to not be confused by single-quotes.

Before this commit, having a single quote anywhere in the format would mess
up all of the format parsing.

Thanks to MinSu Seo for reporting.

11 years agoeo: little micro optimization.
Cedric Bail [Mon, 1 Jul 2013 09:35:55 +0000 (18:35 +0900)]
eo: little micro optimization.

11 years agoeo: a little more inlining, give me a 10% speed improvement.
Cedric Bail [Mon, 1 Jul 2013 08:09:02 +0000 (17:09 +0900)]
eo: a little more inlining, give me a 10% speed improvement.

11 years agoRevert "configure: Link eina to librt. This is needed for shm_* calls"
Daniel Willmann [Mon, 1 Jul 2013 08:18:25 +0000 (09:18 +0100)]
Revert "configure: Link eina to librt. This is needed for shm_* calls"

This reverts commit 6445fea318f29fd6b1de0bc6fbe1b66bcc5673ba.

Jenkins seems lost without it, but this is wrong on some platforms. Need
to solve it another way.

11 years ago* Evas: Fix jpeg loader cannot deal with exif information correctly
Jiyoun Park [Sat, 29 Jun 2013 06:00:43 +0000 (15:00 +0900)]
* Evas: Fix jpeg loader cannot deal with exif information correctly

11 years agoedje_test: Test geometry of a more complex layout
Daniel Willmann [Thu, 27 Jun 2013 16:53:48 +0000 (17:53 +0100)]
edje_test: Test geometry of a more complex layout

It seems someone thought it was fun to have rel2.offset behave
differently than rel1.offset...
A rel2.offset of 0 actually means the size is increased by one pixel.

Acocunt for that in the tests.

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoedje_test: Make the variable returned in test_layout_get static
Daniel Willmann [Fri, 28 Jun 2013 10:04:21 +0000 (11:04 +0100)]
edje_test: Make the variable returned in test_layout_get static

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoFix klockwork possible memleak report:
Chris Michael [Fri, 28 Jun 2013 13:23:57 +0000 (14:23 +0100)]
Fix klockwork possible memleak report:

If we get an error trying to "make current" an allocated surface, then
we should free that allocated surface.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoFix klockwork memleak: We should free 'texts' that we malloc'd earlier
Chris Michael [Fri, 28 Jun 2013 13:14:44 +0000 (14:14 +0100)]
Fix klockwork memleak: We should free 'texts' that we malloc'd earlier
if we are going to error out.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoWe should free glyph_out also (as that gets malloc'd too) on error.
Chris Michael [Fri, 28 Jun 2013 13:13:09 +0000 (14:13 +0100)]
We should free glyph_out also (as that gets malloc'd too) on error.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoFix klockwork memleak: If we allocate space for a Font Glyph (through
Chris Michael [Fri, 28 Jun 2013 13:02:00 +0000 (14:02 +0100)]
Fix klockwork memleak: If we allocate space for a Font Glyph (through
evas_common_font_int_cache_glyph_get), then we should free that
allocation when we have an error.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoFix klockwork memleak: If we allocate a colormap above, then we need
Chris Michael [Fri, 28 Jun 2013 12:57:56 +0000 (13:57 +0100)]
Fix klockwork memleak: If we allocate a colormap above, then we need
to free it on error.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoFix klockwork error: Null pointer may be dereferenced
Chris Michael [Fri, 28 Jun 2013 12:48:21 +0000 (13:48 +0100)]
Fix klockwork error: Null pointer may be dereferenced

Check for valid out_file before trying to use it.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoFix memleak reported by klockwork:
Chris Michael [Fri, 28 Jun 2013 12:39:45 +0000 (13:39 +0100)]
Fix memleak reported by klockwork:

If a jpeg image is rotated and we have allocated space for ptr_rotate,
then we should free it Always ... Not Only if the degrees are
180...because we allocate it always

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoevas: efl 1.8 dev bug only... update add bug in image objects throws away some updates.
Carsten Haitzler (Rasterman) [Fri, 28 Jun 2013 08:25:30 +0000 (17:25 +0900)]
evas: efl 1.8 dev bug only... update add bug in image objects throws away some updates.

11 years agoupdate po's
Carsten Haitzler (Rasterman) [Fri, 28 Jun 2013 08:01:42 +0000 (17:01 +0900)]
update po's

11 years agoeo: add a new macro to make the code easier to read and use it.
Cedric Bail [Fri, 28 Jun 2013 06:28:30 +0000 (15:28 +0900)]
eo: add a new macro to make the code easier to read and use it.

11 years agoeo: 30% speed improvement in message propagation.
Cedric Bail [Thu, 27 Jun 2013 08:58:49 +0000 (17:58 +0900)]
eo: 30% speed improvement in message propagation.

11 years agoeo: eo is absolutely necessary by now, can't be disabled any more.
Cedric Bail [Thu, 27 Jun 2013 06:51:07 +0000 (15:51 +0900)]
eo: eo is absolutely necessary by now, can't be disabled any more.

11 years agoevas: prevent people from including header directly.
Cedric Bail [Thu, 27 Jun 2013 06:27:30 +0000 (15:27 +0900)]
evas: prevent people from including header directly.

11 years agoeldbus: Missing dbus_connection_unref() in error path
Eduardo Lima (Etrunko) [Thu, 27 Jun 2013 20:31:22 +0000 (17:31 -0300)]
eldbus: Missing dbus_connection_unref() in error path

Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
11 years agoeldbus: Call dbus_bus_register() for TYPE_ADDRESS connections
Eduardo Lima (Etrunko) [Thu, 27 Jun 2013 20:24:46 +0000 (17:24 -0300)]
eldbus: Call dbus_bus_register() for TYPE_ADDRESS connections

From D-Bus documentation:

http://dbus.freedesktop.org/doc/api/html/group__DBusBus.html

dbus_bus_register():

    If you open a bus connection with dbus_connection_open() or
    dbus_connection_open_private() you will have to dbus_bus_register()
    yourself, or make the appropriate registration method calls yourself.

Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
11 years agoeldbus: Add proxy call to org.freedesktop.DBus.Hello
Eduardo Lima (Etrunko) [Thu, 27 Jun 2013 19:38:53 +0000 (16:38 -0300)]
eldbus: Add proxy call to org.freedesktop.DBus.Hello

Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
11 years agoedje_test: Test geometry and color of edje part
Daniel Willmann [Thu, 27 Jun 2013 16:51:30 +0000 (17:51 +0100)]
edje_test: Test geometry and color of edje part

Just a simple test that uses the existing test_layout

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoedje_test: Make test_layout_get() configurable
Daniel Willmann [Thu, 27 Jun 2013 16:49:56 +0000 (17:49 +0100)]
edje_test: Make test_layout_get() configurable

This will help when we want to test different edj files

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoeina_test_fp: Improve speed of eina_fp test
Daniel Willmann [Thu, 27 Jun 2013 13:27:36 +0000 (14:27 +0100)]
eina_test_fp: Improve speed of eina_fp test

Test less values in between, but still have a decent combination of
large/small values to test.

Fixes -WTasn

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoWhen getting crtc info, fill in missing "rotations" field so that the
Chris Michael [Thu, 27 Jun 2013 09:45:04 +0000 (10:45 +0100)]
When getting crtc info, fill in missing "rotations" field so that the
value is initialized.

NB: Spank Spank whoever did this ... oh wait, that was me :( Ouch !!

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoedje_entry: Fix memory leak in _edje_entry_imf_event_preedit_changed_cb
Stefan Schmidt [Thu, 27 Jun 2013 08:43:08 +0000 (09:43 +0100)]
edje_entry: Fix memory leak in _edje_entry_imf_event_preedit_changed_cb

info is only local and we are going out of scope after allocating memory
for it. As we are going out of scope anyway we can safely free it before.

11 years agoefreet: always sanitize path.
Cedric Bail [Thu, 27 Jun 2013 03:25:03 +0000 (12:25 +0900)]
efreet: always sanitize path.

11 years agoefreet: cleanup code by using Eina_File infra.
Cedric Bail [Thu, 27 Jun 2013 03:23:33 +0000 (12:23 +0900)]
efreet: cleanup code by using Eina_File infra.

11 years agoevas_convert_colorspace: Remove printf in e_c_convert_yuv_422_601_to()
Daniel Willmann [Wed, 26 Jun 2013 17:41:32 +0000 (18:41 +0100)]
evas_convert_colorspace: Remove printf in e_c_convert_yuv_422_601_to()

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoevas_convert_colorspace: Be nice and actually return the converted data
Daniel Willmann [Wed, 26 Jun 2013 17:38:24 +0000 (18:38 +0100)]
evas_convert_colorspace: Be nice and actually return the converted data

Fixes a memory leaks in evas_common_convert_yuv_42* and actually makes
these functions more useful. It's a win-win.

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoecore_wayland: Yes, we need to call display_dispatch
Eduardo Lima (Etrunko) [Wed, 26 Jun 2013 14:35:59 +0000 (11:35 -0300)]
ecore_wayland: Yes, we need to call display_dispatch

Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
11 years agoWIP
Eduardo Lima (Etrunko) [Wed, 26 Jun 2013 13:55:14 +0000 (10:55 -0300)]
WIP

Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
11 years agoevas gl_x11: Fix uninitialized use of re in eng_output_flush()
Daniel Willmann [Wed, 26 Jun 2013 13:32:21 +0000 (14:32 +0100)]
evas gl_x11: Fix uninitialized use of re in eng_output_flush()

In case of render_mode == EVAS_RENDER_MODE_ASYNC_INIT we're jumping
directly to end which then calls
evas_gl_preload_render_unlock(eng_preload_make_current, re);

Make sure re is initialized with the render engine data so we don't
dereference junk.

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoevas gl_common: Fix incorrect number of parameters in functions
Daniel Willmann [Wed, 26 Jun 2013 13:30:48 +0000 (14:30 +0100)]
evas gl_common: Fix incorrect number of parameters in functions

_evgl_tls_resource_get() and _surface_context_list_print() don't take
parameters.

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agopixes -> pixels.
Raphael Kubo da Costa [Wed, 26 Jun 2013 13:26:35 +0000 (16:26 +0300)]
pixes -> pixels.

11 years agoAdd missing initializer for wl_seat_listener.
Chris Michael [Wed, 26 Jun 2013 11:39:14 +0000 (12:39 +0100)]
Add missing initializer for wl_seat_listener.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agofeature: Add support for global_remove in the display listener.
Chris Michael [Wed, 26 Jun 2013 10:52:58 +0000 (11:52 +0100)]
feature: Add support for global_remove in the display listener.
bugfix T151: Catch fatal error from wayland displays and signal apps to exit.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoconfigure: Link eina to librt. This is needed for shm_* calls
Daniel Willmann [Wed, 26 Jun 2013 09:57:24 +0000 (10:57 +0100)]
configure: Link eina to librt. This is needed for shm_* calls

Recent clang breaks with linking errors:
lib/eina/.libs/libeina.so: undefined reference to `shm_open'
so fix.

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoedje_cc: Fix memory leak in data_thread_script() in error case
Daniel Willmann [Tue, 25 Jun 2013 15:23:03 +0000 (16:23 +0100)]
edje_cc: Fix memory leak in data_thread_script() in error case

Signed-off-by: Daniel Willmann <d.willmann@samsung.com>
11 years agoEvas gl: Fixed double-free when failing gl init.
Tom Hacohen [Wed, 26 Jun 2013 09:40:44 +0000 (10:40 +0100)]
Evas gl: Fixed double-free when failing gl init.

11 years agoefreet: and dont try change mode if runtime dir is /tmp
Carsten Haitzler (Rasterman) [Wed, 26 Jun 2013 03:16:50 +0000 (12:16 +0900)]
efreet: and dont try change mode if runtime dir is /tmp

11 years agoevas: check wether we do have a data or not.
Cedric Bail [Wed, 26 Jun 2013 02:49:16 +0000 (11:49 +0900)]
evas: check wether we do have a data or not.

11 years agoefreet: make geteuid an optional function.
Cedric Bail [Tue, 25 Jun 2013 09:11:43 +0000 (18:11 +0900)]
efreet: make geteuid an optional function.

11 years agoevas: cleanup headers. Should work on windows.
Cedric Bail [Tue, 25 Jun 2013 07:44:37 +0000 (16:44 +0900)]
evas: cleanup headers. Should work on windows.

11 years agoEcore DPMS Function to get the current dpms power level
Deon Thomas [Tue, 25 Jun 2013 23:32:42 +0000 (08:32 +0900)]
Ecore DPMS Function to get the current dpms power level

11 years agoEvas textblock: Fixed issue with textblocks without fonts segfaulting.
Tom Hacohen [Tue, 25 Jun 2013 15:37:48 +0000 (16:37 +0100)]
Evas textblock: Fixed issue with textblocks without fonts segfaulting.

Fixes T184 (which is a regression).

11 years agoeina_list: Avoid NULL pointer deref in eina_list_shuffle.
Stefan Schmidt [Tue, 25 Jun 2013 14:18:33 +0000 (15:18 +0100)]
eina_list: Avoid NULL pointer deref in eina_list_shuffle.

If we have a valid but empty, n = 0, list we never enter the loop
and shuffled_last might still be NULL. Avoid this case by checking
explicitly for n = 0 before.

11 years agoemotion: Check engine for NULL before dereferencing.
Stefan Schmidt [Tue, 25 Jun 2013 13:36:38 +0000 (14:36 +0100)]
emotion: Check engine for NULL before dereferencing.

It might be that our engine list is empty and thus our engine might still
be set to NULL. Better check and return if this is the case instead of
dereferencing NULL.

11 years agoFormat for EFL style
Chris Michael [Tue, 25 Jun 2013 13:01:30 +0000 (14:01 +0100)]
Format for EFL style

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoAdd missing safety checks for valid proxy objects before trying to
Chris Michael [Tue, 25 Jun 2013 12:51:53 +0000 (13:51 +0100)]
Add missing safety checks for valid proxy objects before trying to
call functions on the object

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoefl/evas - removed duplicated lines and apply clipping only when clip values are...
ChunEon Park [Tue, 25 Jun 2013 12:56:28 +0000 (21:56 +0900)]
efl/evas - removed duplicated lines and apply clipping only when clip values are available.

11 years agoefl/evas - +null check
ChunEon Park [Tue, 25 Jun 2013 12:53:05 +0000 (21:53 +0900)]
efl/evas - +null check

11 years agoCheck that ecore_wl_init succeeds before trying to make calls into it.
Chris Michael [Tue, 25 Jun 2013 08:31:52 +0000 (09:31 +0100)]
Check that ecore_wl_init succeeds before trying to make calls into it.
Fix formatting.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoevas: don't forget to remove image from the lru when they are back to active.
Cedric Bail [Tue, 25 Jun 2013 07:20:46 +0000 (16:20 +0900)]
evas: don't forget to remove image from the lru when they are back to active.

11 years agoevas: fix problem with dllimport on Windows.
Christophe Sadoine [Tue, 25 Jun 2013 03:27:45 +0000 (12:27 +0900)]
evas: fix problem with dllimport on Windows.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevil: add strsep function.
Christophe Sadoine [Tue, 25 Jun 2013 03:26:20 +0000 (12:26 +0900)]
evil: add strsep function.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevil: added evil_rename() a wrapper around rename().
Christophe Sadoine [Wed, 19 Jun 2013 06:47:20 +0000 (15:47 +0900)]
evil: added evil_rename() a wrapper around rename().

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevil: returned the good value for mkstemp
Cedric Bail [Tue, 25 Jun 2013 03:16:36 +0000 (12:16 +0900)]
evil: returned the good value for mkstemp

Added back some needed defines for elementary (was in evil 1.7.7).

11 years agoeet: move the dependency for crypto.
Cedric Bail [Tue, 25 Jun 2013 03:16:03 +0000 (12:16 +0900)]
eet: move the dependency for crypto.

11 years agoedje: leak is gone, so use it again.
Cedric Bail [Tue, 25 Jun 2013 02:44:38 +0000 (11:44 +0900)]
edje: leak is gone, so use it again.

11 years agoevas: one LRU for both image open with filename or Eina_File to fix infinite increase...
Cedric Bail [Tue, 25 Jun 2013 02:42:14 +0000 (11:42 +0900)]
evas: one LRU for both image open with filename or Eina_File to fix infinite increase of cache.

11 years agoedje: fix memory leak with EDJE_TWEEN_MODE_OPT_FROM_CURRENT.
Ryuan Choi [Tue, 25 Jun 2013 01:37:59 +0000 (10:37 +0900)]
edje: fix memory leak with EDJE_TWEEN_MODE_OPT_FROM_CURRENT.

11 years agoefreet: Init no longer fails without dbus session
Sebastian Dransfeld [Tue, 25 Jun 2013 02:32:26 +0000 (04:32 +0200)]
efreet: Init no longer fails without dbus session

Allow programs to use efreet without requiering a dbus session. This
gives limited functionality, as efreet_icon wont work without a cache.
efreet_desktop will partially work, as it reads info from files directly
if cache is missing.

11 years agoecore_imf: Add support for wayland
Jan Arne Petersen [Wed, 17 Apr 2013 09:59:46 +0000 (11:59 +0200)]
ecore_imf: Add support for wayland

Add an input method module supporting the Wayland input method protocol.

In v2:
 - Missing call to ecore_wl_shutdown() in im_module_shutdown()
 - Access the already stored 'wayland globals' with ecore_wl_globals_get()
 - Fixed compilation by adding -I(top_srcdir)/src/efl to CFLAGS
 - EFL formatting fixes

Signed-off-by: Eduardo Lima (Etrunko) <eduardo.lima@intel.com>
11 years agoevas: fix crash after font rehint
Jean-Philippe Andre [Mon, 24 Jun 2013 12:28:27 +0000 (14:28 +0200)]
evas: fix crash after font rehint

11 years agoPort cedric's async uploading GL texture commit to wayland_egl engine.
Chris Michael [Mon, 24 Jun 2013 09:14:21 +0000 (10:14 +0100)]
Port cedric's async uploading GL texture commit to wayland_egl engine.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoMake gl_x11 compile again
Chris Michael [Mon, 24 Jun 2013 08:27:14 +0000 (09:27 +0100)]
Make gl_x11 compile again
eglSwap|BuffersRegion was replaced with eglSwapBuffersWithDamage.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoRemove call to wl_display_get_serial as it's a server-side function,
Chris Michael [Mon, 24 Jun 2013 08:15:34 +0000 (09:15 +0100)]
Remove call to wl_display_get_serial as it's a server-side function,
not client-side.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoUse wl_display_get_serial when creating popups.
Chris Michael [Fri, 14 Jun 2013 08:21:43 +0000 (09:21 +0100)]
Use wl_display_get_serial when creating popups.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
11 years agoefl: fix systemd spelling.
Doug Newgard [Mon, 24 Jun 2013 02:52:57 +0000 (11:52 +0900)]
efl: fix systemd spelling.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas: add mutex locking before signalling for the condition variable and releasing...
Rajeev Ranjan [Mon, 24 Jun 2013 02:50:08 +0000 (11:50 +0900)]
evas: add mutex locking before signalling for the condition variable and releasing afterwards.

Signed-off-by: Cedric Bail <cedric.bail@samsung.com>
11 years agoevas: add support for asynchronously uploading GL texture.
Cedric Bail [Mon, 24 Jun 2013 02:41:32 +0000 (11:41 +0900)]
evas: add support for asynchronously uploading GL texture.

NOTE: when using Evas_Object image preload infrastructure the GL texture
upload was uploaded from the main loop during the rendering stage. This
could lead to some frame drop during fast animation due to the time needed
to upload that texture.

This patch fix this problem by uploading a small texture quickly (16x16)
and waiting for going back to the main loop to be able to use the same GL
context from another thread to do the texture upload asynchronously without
blocking the main loop.

11 years agoefreet: don't try change permissions on runtime dir we don't own.
Carsten Haitzler (Rasterman) [Mon, 24 Jun 2013 03:00:30 +0000 (12:00 +0900)]
efreet: don't try change permissions on runtime dir we don't own.