platform/upstream/efl.git
10 years agoecore_imf_example: remove wrong implementation code.
Jihoon Kim [Sat, 4 Jan 2014 05:23:38 +0000 (14:23 +0900)]
ecore_imf_example: remove wrong implementation code.

Enter or Escape key should be processed by input method framework.
Editor should not call ecore_imf_context_reset by itself.

10 years agoUse key instead of keyname for considering setxkbmap
Jihoon Kim [Sat, 4 Jan 2014 04:25:21 +0000 (13:25 +0900)]
Use key instead of keyname for considering setxkbmap

10 years agoMajor cleanup of the Emotion GStreamer 1.x video sink
Sebastian Dröge [Sat, 4 Jan 2014 02:41:37 +0000 (11:41 +0900)]
Major cleanup of the Emotion GStreamer 1.x video sink

Next step: state handling in the GStreamer backend

Reviewers: cedric

CC: cedric
Differential Revision: https://phab.enlightenment.org/D431

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoeina: fix support for system requiring _XOPEN_SOURCE to use spinlock and barrier.
Cedric Bail [Sat, 4 Jan 2014 02:25:19 +0000 (11:25 +0900)]
eina: fix support for system requiring _XOPEN_SOURCE to use spinlock and barrier.

10 years agoeina: fix build break on systemd based system.
Cedric Bail [Sat, 4 Jan 2014 02:24:48 +0000 (11:24 +0900)]
eina: fix build break on systemd based system.

10 years agoecore_evas: extn - restore missing symbol that resulted in a temporary ABI break.
Cedric Bail [Sat, 4 Jan 2014 02:23:47 +0000 (11:23 +0900)]
ecore_evas: extn - restore missing symbol that resulted in a temporary ABI break.

Thanks to Albin and Debian tools to have spotted that.

- cherry-pick me -

10 years agoeina - Fix compilation warning in eina_log_print_cb_journald.
Guillaume Friloux [Fri, 3 Jan 2014 18:13:30 +0000 (19:13 +0100)]
eina - Fix compilation warning in eina_log_print_cb_journald.

vasprintf can return -1, in which case the buffer is corrupted.
So we better handle this ...

gcc told me of this thanks to -Wunused-result when building package,
thank you gcc for your incredible powers.

10 years agoeina - Fix coding style madness.
Guillaume Friloux [Fri, 3 Jan 2014 18:10:16 +0000 (19:10 +0100)]
eina - Fix coding style madness.

10 years agoecore_audio - Fix doxygen command that does not exist.
Guillaume Friloux [Fri, 3 Jan 2014 18:07:03 +0000 (19:07 +0100)]
ecore_audio - Fix doxygen command that does not exist.

10 years agoecore_con - Fix binary buffer when using ECORE_CON_REMOTE_UDP.
Nicolas Aguirre [Fri, 3 Jan 2014 17:57:56 +0000 (18:57 +0100)]
ecore_con - Fix binary buffer when using ECORE_CON_REMOTE_UDP.

ERR<7807>: lib/eina/eina_binbuf_template_c.x:95 eina_binbuf_append_length() *** Eina Magic Check Failed !!!
This fix a problem where eina_binbuf was used without
calling eina_binbuf_new when ECORE_CON_REMOTE_UDP is used.

10 years agoUpdating serbian and italian translations
maxerba [Fri, 3 Jan 2014 17:18:55 +0000 (18:18 +0100)]
Updating serbian and italian translations

10 years agoevas: disable use of mmap on windows for allocating pixels buffer.
Cedric BAIL [Fri, 3 Jan 2014 08:53:51 +0000 (17:53 +0900)]
evas: disable use of mmap on windows for allocating pixels buffer.

We do have mmap provided by Evil, but there is no implementation yet of
an anonymous map support. Also it is not clear how the memory system of
windows does actually work, so not sure this optimization is relevant
to windows at all. Thus we disable it for the time being and unbreak
the windows support.

- cherry-pick me -

10 years agoscimimmodule: remove unused code
Jihoon Kim [Fri, 3 Jan 2014 07:23:50 +0000 (16:23 +0900)]
scimimmodule: remove unused code

need_commit_preedit was always false, so those code are useless.

10 years agoemotion: Fix missing include resulting in unknown type GstNavigationCommand
Stefan Schmidt [Thu, 2 Jan 2014 12:54:39 +0000 (13:54 +0100)]
emotion: Fix missing include resulting in unknown type GstNavigationCommand

At least on the gstreamer1 version in Fedora 19 this include is needed. Glima
reported it as well and I think he also uses Fedora.

modules/emotion/gstreamer1/emotion_gstreamer.c:643:4: error: unknown type name
'GstNavigationCommand'

Even if other distros or gstreamer1 versions do not need this it should be safe
to add it here.

10 years agoeina: make eina_binbuf_free and eina_strbuff_free accept NULL.
Cedric BAIL [Thu, 2 Jan 2014 03:11:38 +0000 (12:11 +0900)]
eina: make eina_binbuf_free and eina_strbuff_free accept NULL.

We do follow free() logic here and accept NULL for all eina_*_free() functions.
This patch just do so for Eina_Binbuf and Eina_Strbuf.

10 years agoscimimmodule: support super key (between ctrl and alt key)
Jihoon Kim [Thu, 2 Jan 2014 06:08:01 +0000 (15:08 +0900)]
scimimmodule: support super key (between ctrl and alt key)

10 years agoscimimmodule: code refinement related to key modifier and lock
Jihoon Kim [Thu, 2 Jan 2014 06:00:51 +0000 (15:00 +0900)]
scimimmodule: code refinement related to key modifier and lock

10 years agofix map-across-mark patch from sergey in december 2013
Carsten Haitzler (Rasterman) [Thu, 2 Jan 2014 00:26:34 +0000 (09:26 +0900)]
fix map-across-mark patch from sergey in december 2013

stable release - cherry-pick me!

of course! eina_hash_direct_add() for the object pointer is using the
poitner to the stack value, not the value itself it points to... this
was bad and just by luck out value was on the stack that grows but
never shrinks and thus never crashes, BUT... it will just break in all
sorts of fun ways. basically it makes the hash useless as the keys in
it are effectively all the SAME value as they point to the same
storage.. but it changes whenever that stack mem gets changed.

10 years agoFix documentation for efreet_desktop_exec().
Guillaume Friloux [Wed, 1 Jan 2014 20:39:14 +0000 (21:39 +0100)]
Fix documentation for efreet_desktop_exec().

10 years agoFix documentation for efreet_desktop_environment_get().
Guillaume Friloux [Wed, 1 Jan 2014 20:28:37 +0000 (21:28 +0100)]
Fix documentation for efreet_desktop_environment_get().

10 years agoSimplify code of efreet_desktop_environment_set().
Guillaume Friloux [Wed, 1 Jan 2014 20:27:19 +0000 (21:27 +0100)]
Simplify code of efreet_desktop_environment_set().

10 years agoevas - sw render - map render neon broken for expedite 45/46 disable
Carsten Haitzler (Rasterman) [Wed, 1 Jan 2014 03:56:18 +0000 (12:56 +0900)]
evas - sw render - map render neon broken for expedite 45/46 disable

10 years agoevas - fix neon scale code
Nick Reed [Wed, 1 Jan 2014 02:22:39 +0000 (11:22 +0900)]
evas - fix neon scale code

Summary: evas_scale_smooth would not compile with BUILD_NEON set

Reviewers: raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D424

10 years agoevas/common - removed white spaces and use eina bool instead of int.
ChunEon Park [Tue, 31 Dec 2013 12:11:58 +0000 (21:11 +0900)]
evas/common - removed white spaces and use eina bool instead of int.

10 years agoevas/common - refactoring code.
ChunEon Park [Tue, 31 Dec 2013 09:32:30 +0000 (18:32 +0900)]
evas/common - refactoring code.

removed white spaces.
added some comments.
changed function and variable name to be more descriptive.

10 years agoevas/common - more elaborated compuatation in interpolation.
ChunEon Park [Tue, 31 Dec 2013 09:19:41 +0000 (18:19 +0900)]
evas/common - more elaborated compuatation in interpolation.

we should not +1 in divide but only do it when the quotient is zero.

10 years agoibusimmodule: support CAPS LOCK and NUM LOCK
Jihoon Kim [Tue, 31 Dec 2013 09:07:36 +0000 (18:07 +0900)]
ibusimmodule: support CAPS LOCK and NUM LOCK

When ibus immodule was used, caps lock and num lock had been not working.

10 years agoevas/common - removed white spaces
ChunEon Park [Tue, 31 Dec 2013 04:29:57 +0000 (13:29 +0900)]
evas/common - removed white spaces

10 years agoevas/common - fixed a comment typo.
ChunEon Park [Tue, 31 Dec 2013 04:28:57 +0000 (13:28 +0900)]
evas/common - fixed a comment typo.

10 years agoibusimmodule: fix segmentation fault
Jihoon Kim [Tue, 31 Dec 2013 01:01:54 +0000 (10:01 +0900)]
ibusimmodule: fix segmentation fault

This patch fixes "free(): invalid pointer issue"

10 years agoedje - fix edje lua usage for lua 5.2 to not go crazy on memory allocs
Carsten Haitzler (Rasterman) [Tue, 31 Dec 2013 01:00:17 +0000 (10:00 +0900)]
edje - fix edje lua usage for lua 5.2 to not go crazy on memory allocs

this fixes T323 - this is a change in lua 5.2 that makes osize encode
the type of data being allocated where 5.1 didn't do it and the math
was thus screwed as a result. comments in the diff.

cherry-pick me!

10 years agoevas - gl-x11 remove +x bit on src file!
Carsten Haitzler (Rasterman) [Mon, 30 Dec 2013 09:27:06 +0000 (18:27 +0900)]
evas - gl-x11 remove +x bit on src file!

10 years agoevas - set clipees list to NULL on free to avoid access
Carsten Haitzler (Rasterman) [Mon, 30 Dec 2013 06:54:21 +0000 (15:54 +0900)]
evas - set clipees list to NULL on free to avoid access

10 years agoecore-evas - x11 - call resize callback if wm denies resize request
Carsten Haitzler (Rasterman) [Mon, 30 Dec 2013 04:39:18 +0000 (13:39 +0900)]
ecore-evas - x11 - call resize callback if wm denies resize request

in SOME cases a wm may deny a resize request entirely and keep the
window size the same. it does so with a configurenotify of the SAME
size as before. problem is we never passed this onto the app with a
resize callback of the original size before request to the new size.
this does that now.

cherry-pick me!

10 years agoconfigure xcb warning typos
discomfitor [Mon, 30 Dec 2013 00:43:39 +0000 (19:43 -0500)]
configure xcb warning typos

10 years agoconfigure - add warning when enabling xcb
Carsten Haitzler (Rasterman) [Mon, 30 Dec 2013 00:14:50 +0000 (09:14 +0900)]
configure - add warning when enabling xcb

10 years agotextblock: Simple refactoring for _markup_get_text_utf8_append to call eina_unicode_u...
Ryuan Choi [Sun, 29 Dec 2013 04:33:19 +0000 (13:33 +0900)]
textblock: Simple refactoring for _markup_get_text_utf8_append to call eina_unicode_utf8_next_get directly

Summary:
evas_string_char_next_get is simple wrapper of eina_unicode_utf8_next_get with
some check routines.
But in _markup_get_text_utf8_append, these check routines look unnecessary.

Reviewers: cedric, seoz, raster

Reviewed By: raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D420

10 years agoBlending function rework and speedup
Yury Usishchev [Sun, 29 Dec 2013 04:27:54 +0000 (13:27 +0900)]
Blending function rework and speedup

Summary:
_op_blend_mas_c_dp_neon rework:
main loop now process 4 pixels per iteration
fast path when *m == 0

Reviewers: raster

Reviewed By: raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D418

10 years agoemotion - generic - set play after setting mutes and play speed etc.
Carsten Haitzler (Rasterman) [Sat, 28 Dec 2013 02:28:51 +0000 (11:28 +0900)]
emotion - generic - set play after setting mutes and play speed etc.

10 years agoethumbd - stop using an idler that spins at 100% cpu while generating queues
Carsten Haitzler (Rasterman) [Sat, 28 Dec 2013 02:15:26 +0000 (11:15 +0900)]
ethumbd - stop using an idler that spins at 100% cpu while generating queues

10 years agoethumbd - if thumb gen takes too long don't just quit ethumbd, but skip thumb
Carsten Haitzler (Rasterman) [Sat, 28 Dec 2013 02:07:30 +0000 (11:07 +0900)]
ethumbd - if thumb gen takes too long don't just quit ethumbd, but skip thumb

10 years agoethumbd - drop priority to 19 like efreetd so it lurks in bg as much as it can
Carsten Haitzler (Rasterman) [Sat, 28 Dec 2013 01:28:57 +0000 (10:28 +0900)]
ethumbd - drop priority to 19 like efreetd so it lurks in bg as much as it can

10 years agoefreet - minor formatting and move set prirority to very start
Carsten Haitzler (Rasterman) [Sat, 28 Dec 2013 01:22:23 +0000 (10:22 +0900)]
efreet - minor formatting and move set prirority to very start

10 years agoEvas textblock: Fixed range_text_get bug with multiple text nodes.
Youngbok Shin [Fri, 27 Dec 2013 11:36:10 +0000 (11:36 +0000)]
Evas textblock: Fixed range_text_get bug with multiple text nodes.

Summary:
When there is multi text nodes for range text get,
it gets wrong format node of last text node.
It makes broken result.

Test Plan: https://phab.enlightenment.org/D398

Reviewers: woohyun, tasn, seoz

Reviewed By: tasn

CC: cedric
Differential Revision: https://phab.enlightenment.org/D399

10 years agoemotion: don't duplicate storage of audio/video stream metadata in GStreamer 1.x...
Sebastian Dröge [Fri, 27 Dec 2013 08:45:56 +0000 (17:45 +0900)]
emotion: don't duplicate storage of audio/video stream metadata in GStreamer 1.x backend.

Reviewers: cedric

CC: cedric
Differential Revision: https://phab.enlightenment.org/D415

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoevas - gl engine - glerr should be off by default for performance
Carsten Haitzler (Rasterman) [Fri, 27 Dec 2013 08:32:15 +0000 (17:32 +0900)]
evas - gl engine - glerr should be off by default for performance

stable release - cherry-pick me!

10 years agooops - and a printf crept in. i thought i was done.
Carsten Haitzler (Rasterman) [Fri, 27 Dec 2013 08:18:46 +0000 (17:18 +0900)]
oops - and a printf crept in. i thought i was done.

10 years agoemotion - broken module loading loagic fixed that fixes gst 0.10/1 deadlock
Carsten Haitzler (Rasterman) [Fri, 27 Dec 2013 08:14:02 +0000 (17:14 +0900)]
emotion - broken module loading loagic fixed that fixes gst 0.10/1 deadlock

so firstly the module loading logic in emotion is pretty bad. it
forcible loads into memory (dlopen + run code from) EVERY emotion
module whenever you use emotion anywhere. this is a fat memory cost and
startup cost. it should not ever have done this. so remove that code
and make it explicitly load only the backend requested and fall back
to using what is compiled in (generic by default) and otherwise
generic as a module, then xine, then gstreamer then gstreamer1.
gstreamer1 seems broke - all i see is a black box (no video).

this also fixes a deadlock problem. if you have BOTH gstreamer AND
gstreamer1 modules loaded i get a deadlock inside glib. this seemingly
fixes it as it'll only load the first one it finds, not both (unless
explicitly requested).

10 years agoEdje: add edje_object_part_text_input_panel_show_on_demand_set/get()
Jihoon Kim [Fri, 27 Dec 2013 04:24:54 +0000 (13:24 +0900)]
Edje: add edje_object_part_text_input_panel_show_on_demand_set/get()

10 years agoeina: struct _Eina_Rbtree contains an enum bitfield, which is a GNU extension not...
Dennis Schridde [Fri, 27 Dec 2013 03:06:30 +0000 (12:06 +0900)]
eina: struct _Eina_Rbtree contains an enum bitfield, which is a GNU extension not specified by C99.

Reviewers: barbieri, raster, cedric

Reviewed By: cedric

CC: cedric
Differential Revision: https://phab.enlightenment.org/D412

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoecore-x - rename new api
Carsten Haitzler (Rasterman) [Fri, 27 Dec 2013 03:02:48 +0000 (12:02 +0900)]
ecore-x - rename new api

10 years agoemotion: remove (broken) X11 integration from the GStreamer 1.x backend.
Sebastian Dröge [Fri, 27 Dec 2013 00:56:11 +0000 (09:56 +0900)]
emotion: remove (broken) X11 integration from the GStreamer 1.x backend.

This code was disabled and did not work properly. If this should be
re-added it will have to be rewritten anyway and currently it only
clutters the existing code.

Reviewers: cedric

CC: cedric
Differential Revision: https://phab.enlightenment.org/D414

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoevas: improve textblock rendering by walking even less in all case.
Cedric BAIL [Fri, 27 Dec 2013 00:34:29 +0000 (09:34 +0900)]
evas: improve textblock rendering by walking even less in all case.

We now build a list of item to walk on for each step to just avoid all necessary walking.
It is a slightly more elegant idea than my previous patch even if it only give a speed
improvement of 5% in the best case. Now that render code and the callee only take 1.6%
of the time I am looking at in the benchmark, meaning nothing else to improve here.

This fix the bug spotted in Enlightenment dialog box.

10 years agoRevert "evas: improve textblock rendering by walking even less in all case."
Carsten Haitzler (Rasterman) [Thu, 26 Dec 2013 15:05:11 +0000 (00:05 +0900)]
Revert "evas: improve textblock rendering by walking even less in all case."

This reverts commit a69c5ba0aee81e84e2f03475faca21bb1c939843.

yes - this broke text rendering. revert it. several dialogs/uses in e
broke with shadow and glow text being heavily offset up/to the right
of the proper text.

10 years agoevas gl engine - ensure on texture delete we never leave a cur tex id dangling
Carsten Haitzler (Rasterman) [Thu, 26 Dec 2013 12:03:56 +0000 (21:03 +0900)]
evas gl engine - ensure on texture delete we never leave a cur tex id dangling

10 years agoCleanup ecore_x_window_permanent_create function:
Chris Michael [Thu, 26 Dec 2013 11:40:32 +0000 (11:40 +0000)]
Cleanup ecore_x_window_permanent_create function:

  - No need for a separate xcb_connection here as we already have a
connection at this point so let's use that.
  - Fix up formatting.

Signed-off-by: Chris Michael <devilhorns@comcast.net>
10 years agoevas: improve textblock rendering by walking even less in all case.
Cedric BAIL [Thu, 26 Dec 2013 10:17:46 +0000 (19:17 +0900)]
evas: improve textblock rendering by walking even less in all case.

We now build a list of item to walk on for each step to just avoid all necessary walking.
It is a slightly more elegant idea than my previous patch even if it only give a speed
improvement of 5% in the best case. Now that render code and the callee only take 1.6%
of the time I am looking at in the benchmark, meaning nothing else to improve here.

10 years agoevas - gl-x11 engine - don't re-init engine window on rotate
Vostokov Sergey [Thu, 26 Dec 2013 10:19:15 +0000 (19:19 +0900)]
evas - gl-x11 engine - don't re-init engine window on rotate

this puts rotationc hanges in the same class as resizes for evas gl
x11 engine info updates and doesnt re-create the engine window data on
rotate.

10 years agoecore-x - add call to create a permanent window that is not deleted on exit
Carsten Haitzler (Rasterman) [Thu, 26 Dec 2013 10:01:23 +0000 (19:01 +0900)]
ecore-x - add call to create a permanent window that is not deleted on exit

10 years agoevas - protect against table cache being null and then used
Carsten Haitzler (Rasterman) [Thu, 26 Dec 2013 03:46:26 +0000 (12:46 +0900)]
evas - protect against table cache being null and then used

10 years agoefl: Unified eina critical manro to CRI.
Daniel Juyung Seo [Thu, 26 Dec 2013 03:22:05 +0000 (12:22 +0900)]
efl: Unified eina critical manro to CRI.

Being annoyed by different types of eina critical macros - CRI, CRIT,
 CRITICAL -, I concluded to unify them to one. Discussed on IRC and
 finally, CRI was chosen to meet the consistency with other macros -
 ERR, WRN, INF, DBG - in terms of the number of characters.
If there is any missing bits, please let me know.

10 years agoedje_convert.c: fixed formatting before working on the file.
Daniel Juyung Seo [Tue, 24 Dec 2013 03:47:41 +0000 (12:47 +0900)]
edje_convert.c: fixed formatting before working on the file.

10 years agoevas: limit how many times we walk over items.
Cedric BAIL [Thu, 26 Dec 2013 02:27:09 +0000 (11:27 +0900)]
evas: limit how many times we walk over items.

This does help for some textblock benchmark with a speed increase of 12% and
the one that don't get better don't get slower either, so let's put that in.

10 years agoecore: allow out-of-source build of ecore_xcb_keysym_table.h.
Dennis Schridde [Thu, 26 Dec 2013 01:34:55 +0000 (10:34 +0900)]
ecore: allow out-of-source build of ecore_xcb_keysym_table.h.

Reviewers: zmike, cedric

CC: cedric, zmike
Differential Revision: https://phab.enlightenment.org/D405

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoevas: fix a segfault in _draw_thread_font_draw, caused by the missing initialisation...
Dennis Schridde [Thu, 26 Dec 2013 01:32:40 +0000 (10:32 +0900)]
evas: fix a segfault in _draw_thread_font_draw, caused by the missing initialisation of a struct field

CC: cedric
Differential Revision: https://phab.enlightenment.org/D407

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoAdding hungarian translation
maxerba [Wed, 25 Dec 2013 21:56:42 +0000 (22:56 +0100)]
Adding hungarian translation

10 years agoRevert "Quick and dirty fallback to gst 0.10 if 1.0 is not found"
Carsten Haitzler (Rasterman) [Tue, 24 Dec 2013 14:45:30 +0000 (23:45 +0900)]
Revert "Quick and dirty fallback to gst 0.10 if 1.0 is not found"

This reverts commit 5676eb4ef92f87b406a9a91112513d5368ba8df4.

10 years agoQuick and dirty fallback to gst 0.10 if 1.0 is not found
Iván Briano [Mon, 23 Dec 2013 16:00:14 +0000 (14:00 -0200)]
Quick and dirty fallback to gst 0.10 if 1.0 is not found

10 years agoecore: remove dead free
Sebastian Dransfeld [Mon, 23 Dec 2013 12:39:47 +0000 (13:39 +0100)]
ecore: remove dead free

Since we now free exe on goto restart, exe will always be NULL.

CID 1139955

10 years agopkgbuild - enable scim by default for input methods and sdl is pointless
Carsten Haitzler (Rasterman) [Mon, 23 Dec 2013 04:38:54 +0000 (13:38 +0900)]
pkgbuild - enable scim by default for input methods and sdl is pointless

10 years agopkgbuild - fix deps list for gst 1.0
Carsten Haitzler (Rasterman) [Mon, 23 Dec 2013 04:00:10 +0000 (13:00 +0900)]
pkgbuild - fix deps list for gst 1.0

10 years agoecore: fix shutdown when using system module.
Cedric BAIL [Mon, 23 Dec 2013 02:58:17 +0000 (11:58 +0900)]
ecore: fix shutdown when using system module.

This patch will detect how many more times ecore_init was called
during initialization and use that as a threshold to do a clean shutdown.
It is a necessary evil as we do have ecore module that will initialize
eldbus that will then reinit ecore_init from within ecore_init and without
a chance for the application to act on it.

I also reenable a test to make sure we will catch earlier this kind of issue.

10 years agoecore: fix tests to avoid failure on successful initialization.
Cedric BAIL [Mon, 23 Dec 2013 02:44:52 +0000 (11:44 +0900)]
ecore: fix tests to avoid failure on successful initialization.

This patch is necessary to work around the fact that if you have some
ecore system module installed, they will be initialized and may use another
library like eldbus that will call ecore_init again. So ecore_init will be
different from 1, but still be a success.

My patch does introduce a work around a problem in that situation. Calling
ecore_shutdown wont work at this stage, as we do have module that did initialize
ecore also. And the only way to make ecore_shutdown happen is to unload those
module and that wont happen as they where initialized from inside ecore... lala !

Yes, we never shutdown ecore properly when we have ecore module that use
something like eldbus.

10 years agoecore_audio: update test to not look at ecore_init return to much due to system module.
Cedric BAIL [Mon, 23 Dec 2013 02:44:26 +0000 (11:44 +0900)]
ecore_audio: update test to not look at ecore_init return to much due to system module.

10 years agoeldbus: fix tests to properly take into account the presence of ecore system module.
Cedric BAIL [Mon, 23 Dec 2013 02:42:12 +0000 (11:42 +0900)]
eldbus: fix tests to properly take into account the presence of ecore system module.

Ecore system module may use dbus, so when you init ecore, it may init eldbus a few
more time (depending on the module that are present on your system). This result in
eldbus_init returning a valid init, but different from 1. This patch update the test
to take that into account.

10 years agoeldbus: don't do a double initialization due to ecore module using eldbus now.
Cedric BAIL [Mon, 23 Dec 2013 02:24:49 +0000 (11:24 +0900)]
eldbus: don't do a double initialization due to ecore module using eldbus now.

10 years agoEvas test textblock: Added a test for a broken range_text_get case.
Youngbok Shin [Mon, 23 Dec 2013 02:01:57 +0000 (11:01 +0900)]
Evas test textblock: Added a test for a broken range_text_get case.

Summary: Added a test for range_text_get case on the text that include multi text node.

Reviewers: tasn, woohyun, seoz

CC: cedric
Differential Revision: https://phab.enlightenment.org/D398

10 years agoemotion: initial port of emotion to GStreamer 1.0
Sebastian Dröge [Mon, 23 Dec 2013 01:50:53 +0000 (10:50 +0900)]
emotion: initial port of emotion to GStreamer 1.0

Some commits to port emotion to GStreamer 1.0 and implement some missing features,
clean up code a bit and fix some bugs on the way.

This works as good as the 0.10 code for me now with the emotion examples,
just the Samsung hardware specific code is commented out. This should be
ported by someone who has such hardware, and also in a clean way now that
GStreamer since 1.0 has features to handle all this properly.

There's still a lot of potential to clean things up and fix many bugs, and also
to implement zerocopy rendering. But those are for later if there's actual
interest in this at all.

Commits:

- Update configure checks and ecore example to GStreamer 1.0
- Initial port of emotion to GStreamer 1.0
- Samsung specific code commented out, should be ported by someone
with the hardware.
- Return GST_FLOW_FLUSHING when the sink is unlocked
- Remove unused GSignal from the sink
- Use GstVideoInfo to store the format details inside the sink
- Add support for pixel-aspect-ratio
- Store video format information in GstVideoInfo for the different video streams
- Use GstAudioInfo to store the audio format information
- Remove some unused defines
- Header cleanup
- Implement initial support for GstNavigation interface
- Implement setting of audio/video channel

Reviewers: cedric

CC: cedric
Differential Revision: https://phab.enlightenment.org/D387

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoedje: edje_decc - added -output flag (save to specified dir)
Andrii Kroitor [Mon, 23 Dec 2013 01:48:08 +0000 (10:48 +0900)]
edje: edje_decc - added -output flag (save to specified dir)

Reviewers: cedric, seoz

Reviewed By: cedric

CC: reutskiy.v.v, cedric
Differential Revision: https://phab.enlightenment.org/D402

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoedje: edje_calc - update _edje_part_description_find
Vyacheslav Reutskiy [Mon, 23 Dec 2013 00:51:43 +0000 (09:51 +0900)]
edje: edje_calc - update _edje_part_description_find

Add to fuction prototype new param: Eina_Bool approximation.
If need exact matching state name and value set EINA_FALSE to
'approximate'. In other cases used EINA_TRUE.

Reviewers: cedric, raster, seoz

CC: cedric
Differential Revision: https://phab.enlightenment.org/D400

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoedje: edje_edit - fix .edc file generation
Andrii Kroitor [Mon, 23 Dec 2013 00:47:51 +0000 (09:47 +0900)]
edje: edje_edit - fix .edc file generation

Reviewers: cedric, seoz

Reviewed By: cedric

CC: reutskiy.v.v, cedric
Differential Revision: https://phab.enlightenment.org/D401

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoevas - redraw the map surface if one of the children has been still changed map.
ChunEon Park [Sun, 22 Dec 2013 09:19:40 +0000 (18:19 +0900)]
evas - redraw the map surface if one of the children has been still changed map.

10 years agoemotion_generic: fix permissions, resource leaks and error handling.
Gustavo Sverzut Barbieri [Sat, 21 Dec 2013 19:03:38 +0000 (17:03 -0200)]
emotion_generic: fix permissions, resource leaks and error handling.

 - create permissions as 0700, this should be private to the running user.

 - do shm_unlink() on error, prevents resource leaks;

 - apply close() on error, prevents resource leaks;

 - handle strerror() properly.

 - shm name uses process id, object name and ecore's time, should be
   cleaner and avoid possible clashes

 - ecore_exe do not need to read from slave stdout/err, this removing
   the flags allows messages to come visible at terminal.

10 years agofix doc typo, ernno usage and add missing close().
Gustavo Sverzut Barbieri [Sat, 21 Dec 2013 17:00:30 +0000 (15:00 -0200)]
fix doc typo, ernno usage and add missing close().

 - fix typo, it's slave not slayer... :-D

 - errno: you can't use errno to report error from failed mmap() if
   you do another call that sets errno after that, in that case
   fprintf(). So have a single print so we can use the errno value of
   the proper call.

 - although the slave generic player is not that important if it leaks
   its single shmfd, let's be pedantic and close it. Manpage says it
   is safe to close filedescriptor after mmap() worked.

10 years agobugfix: free the entire ecore_con dns struct on info clear to prevent crashes later on
discomfitor [Sat, 21 Dec 2013 03:16:59 +0000 (22:16 -0500)]
bugfix: free the entire ecore_con dns struct on info clear to prevent crashes later on

10 years agosplit ecore_con_dns free functions into one that removes from list and one that does...
discomfitor [Sat, 21 Dec 2013 03:16:29 +0000 (22:16 -0500)]
split ecore_con_dns free functions into one that removes from list and one that does the freeing

10 years agoRevert "workaround for a weird ecore-con crash"
discomfitor [Sat, 21 Dec 2013 03:04:58 +0000 (22:04 -0500)]
Revert "workaround for a weird ecore-con crash"

This reverts commit c751a220d6bbcb22b120c089e406f278e1891147.

10 years agoworkaround for a weird ecore-con crash
discomfitor [Sat, 21 Dec 2013 02:36:58 +0000 (21:36 -0500)]
workaround for a weird ecore-con crash

kuri ping in case I don't get to this soon it needs to be fixed somewhere else

10 years agofix non-avahi build
Mike Blumenkrantz [Fri, 20 Dec 2013 15:24:13 +0000 (10:24 -0500)]
fix non-avahi build

damn borkers!!!!

10 years agoextn: remove files.{w,h} that are not used.
Gustavo Sverzut Barbieri [Fri, 20 Dec 2013 15:10:38 +0000 (13:10 -0200)]
extn: remove files.{w,h} that are not used.

10 years agoextn: remove unused attribute am_server.
Gustavo Sverzut Barbieri [Fri, 20 Dec 2013 15:09:52 +0000 (13:09 -0200)]
extn: remove unused attribute am_server.

10 years agoextn: complete with NULL newly added functions.
Gustavo Sverzut Barbieri [Fri, 20 Dec 2013 15:08:51 +0000 (13:08 -0200)]
extn: complete with NULL newly added functions.

10 years agoextn: fix typo, alpha is SHM_REF2, not 1!
Gustavo Sverzut Barbieri [Fri, 20 Dec 2013 14:58:45 +0000 (12:58 -0200)]
extn: fix typo, alpha is SHM_REF2, not 1!

10 years agoevas - removed unnecessary doc info
ChunEon Park [Fri, 20 Dec 2013 11:26:57 +0000 (20:26 +0900)]
evas - removed unnecessary doc info

10 years agoevas render2 - more work on basics
Carsten Haitzler (Rasterman) [Fri, 20 Dec 2013 10:45:17 +0000 (19:45 +0900)]
evas render2 - more work on basics

10 years agoevas: fix compilation issue on Jenkins windows that make me wonder how it does build...
Cedric BAIL [Fri, 20 Dec 2013 07:09:10 +0000 (16:09 +0900)]
evas: fix compilation issue on Jenkins windows that make me wonder how it does build on Linux.

10 years agoevas: no loader should ever close an Eina_File anymore.
Cedric BAIL [Fri, 20 Dec 2013 05:39:57 +0000 (14:39 +0900)]
evas: no loader should ever close an Eina_File anymore.

This would lead to some crash in EFM if relying on Evas to find the file format by
trying all its loader.

This should fix T674 and T668.

10 years agoevas: patched evas_map_image.c to use NEON code
Yury Usishchev [Fri, 20 Dec 2013 05:02:48 +0000 (14:02 +0900)]
evas: patched evas_map_image.c to use NEON code

Reviewers: cedric, raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D397

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoedje: edje_cc - note for future image_id changes added
Andrii Kroitor [Thu, 19 Dec 2013 13:51:48 +0000 (22:51 +0900)]
edje: edje_cc - note for future image_id changes added

Reviewers: cedric

Reviewed By: cedric

CC: reutskiy.v.v, cedric
Differential Revision: https://phab.enlightenment.org/D396

Signed-off-by: Cedric BAIL <cedric.bail@samsung.com>
10 years agoedje: edc ref - update the information about fill type values
Vyacheslav Reutskiy [Thu, 19 Dec 2013 13:49:50 +0000 (22:49 +0900)]
edje: edc ref - update the information about fill type values

Reviewers: cedric, raster, seoz

Reviewed By: cedric

CC: cedric
Differential Revision: https://phab.enlightenment.org/D395