platform/upstream/efl.git
10 years agoEolian: fix issue for Windows.
Daniel Zaoui [Wed, 28 May 2014 06:32:56 +0000 (09:32 +0300)]
Eolian: fix issue for Windows.

On Windows, '\' are used to separate the names of a path. It was not
supported.
This patch fixes that. Thanks to Vincent Torri for catching it (and good
luck in Windows ;-))

10 years agogl_x11/evas_engine: Prevent dereference before NULL check
Jaehyun Cho [Wed, 28 May 2014 05:32:39 +0000 (14:32 +0900)]
gl_x11/evas_engine: Prevent dereference before NULL check

Summary: re->win pointer was not compared with NULL pointer before re->win was referenced.

Reviewers: Hermet

Reviewed By: Hermet

CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D910

10 years agoevas_gl_rectangle: Remove unnecessary conditional expression
Jaehyun Cho [Wed, 28 May 2014 05:29:22 +0000 (14:29 +0900)]
evas_gl_rectangle: Remove unnecessary conditional expression

Summary: The comparison gc->dc with NULL is not necessary. So the unnecessary conditional expression is removed.

Reviewers: Hermet

CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D909

10 years agoevas/textblock - apply efl coding style.
ChunEon Park [Wed, 28 May 2014 05:19:46 +0000 (14:19 +0900)]
evas/textblock - apply efl coding style.

10 years agoevas/textblock - removed unnecessary null check.
ChunEon Park [Wed, 28 May 2014 05:14:35 +0000 (14:14 +0900)]
evas/textblock - removed unnecessary null check.

It should be definitely valid.
It's meaningless checking the value there.

10 years agoevas/textblock - null check format.
ChunEon Park [Wed, 28 May 2014 05:07:20 +0000 (14:07 +0900)]
evas/textblock - null check format.

logically it's insane that it doesn't check the null there but it checks right next line.

10 years agoEvas rg_etc2: Fix Windows build
Jean-Philippe Andre [Wed, 28 May 2014 03:04:36 +0000 (12:04 +0900)]
Evas rg_etc2: Fix Windows build

Use unsigned char/int instead of stdint in the header file.

10 years agoevas_draw_main: Remove unnecessary conditional expression
Jaehyun Cho [Wed, 28 May 2014 02:24:56 +0000 (11:24 +0900)]
evas_draw_main: Remove unnecessary conditional expression

Summary: The comparison dc with NULL is not necessary. So remove the unnecessary conditional expression.

Reviewers: Hermet

Reviewed By: Hermet

CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D908

10 years agoevas_font_dir: Prevent the use after free
Jaehyun Cho [Wed, 28 May 2014 02:22:51 +0000 (11:22 +0900)]
evas_font_dir: Prevent the use after free

Summary: Pointer fd is used after being deallocated.

Reviewers: Hermet

Reviewed By: Hermet

CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D906

10 years agoEolian test: Fix other compilation warnings
Jean-Philippe Andre [Wed, 28 May 2014 02:18:03 +0000 (11:18 +0900)]
Eolian test: Fix other compilation warnings

10 years agoEolian test: Fix test on hardened gentoo
Jean-Philippe Andre [Wed, 28 May 2014 02:12:45 +0000 (11:12 +0900)]
Eolian test: Fix test on hardened gentoo

Invalid snprintf, why declare MAX_PATH and then use PATH_MAX?

10 years agoEvas rg_etc1: Fix build break (with DEBUG)
Jean-Philippe Andre [Tue, 27 May 2014 23:12:29 +0000 (08:12 +0900)]
Evas rg_etc1: Fix build break (with DEBUG)

Ooops, bad code merge here.
This breaks the nightly builds.

10 years agoevas/evas3d - fixed incorrect usage in evas-3d-aabb example.
ChunEon Park [Tue, 27 May 2014 13:41:40 +0000 (22:41 +0900)]
evas/evas3d - fixed incorrect usage in evas-3d-aabb example.

10 years agoRevert "ecore/ecore_event - logically this check is redundant."
ChunEon Park [Tue, 27 May 2014 13:24:44 +0000 (22:24 +0900)]
Revert "ecore/ecore_event - logically this check is redundant."

This reverts commit df60e63bf9968a4528577762a7da262466ac7683.

seems this patch is wrong.

10 years agoEvas textblock: Fix wrong advance calculation when bidi is disabled.
Tom Hacohen [Tue, 27 May 2014 13:18:40 +0000 (14:18 +0100)]
Evas textblock: Fix wrong advance calculation when bidi is disabled.

The last_it was not updated correctly with bidi disabled.

Thanks to Daniel Hirt for spotting this.

@fix

10 years agoedje/epp - clean up memory allocated.
ChunEon Park [Tue, 27 May 2014 13:17:01 +0000 (22:17 +0900)]
edje/epp - clean up memory allocated.

10 years ago[evas/evas-3d] Add OBB and AABB.
Oleksandr Shcherbina [Tue, 27 May 2014 13:07:28 +0000 (22:07 +0900)]
[evas/evas-3d] Add OBB and AABB.

Summary: Add axis-aligned bounding box (AABB) to frame. Set and update AABB and oriented bounding box (OBB) in node.

Reviewers: Hermet, cedric, raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D881

10 years agoedje/edje_cc - intensive null check.
ChunEon Park [Tue, 27 May 2014 13:00:59 +0000 (22:00 +0900)]
edje/edje_cc - intensive null check.

10 years agoecore/ecore_exe - removed duplicated null check.
ChunEon Park [Tue, 27 May 2014 11:44:37 +0000 (20:44 +0900)]
ecore/ecore_exe - removed duplicated null check.

10 years agoecore/ecore_event - logically this check is redundant.
ChunEon Park [Tue, 27 May 2014 11:34:42 +0000 (20:34 +0900)]
ecore/ecore_event - logically this check is redundant.

remove check so that prevent tool doens't notify this anymore.

10 years agoautotools: Add xz as additional tarball format
Stefan Schmidt [Tue, 27 May 2014 09:15:35 +0000 (11:15 +0200)]
autotools: Add xz as additional tarball format

If this turns out well we will remove bz2 in 1.12

10 years agoEolian: fix coverity issues.
Daniel Zaoui [Tue, 27 May 2014 05:18:33 +0000 (08:18 +0300)]
Eolian: fix coverity issues.

CID 1216274
CID 1216275
CID 1216276
CID 1216277
CID 1216278
CID 1216279
CID 1216280

10 years agoEvas rg_etc: Show in the doc that we use BGRA now
Jean-Philippe Andre [Thu, 8 May 2014 01:48:35 +0000 (10:48 +0900)]
Evas rg_etc: Show in the doc that we use BGRA now

10 years agoEvas rg_etc2: Small performance fix
Jean-Philippe Andre [Wed, 7 May 2014 08:08:19 +0000 (17:08 +0900)]
Evas rg_etc2: Small performance fix

Alpha is always 255 otherwise we need to loop around and
clamp the values to 255 anyways.

10 years agoEvas rg_etc1: Also decode RGBA8 from the TGV loader
Jean-Philippe Andre [Wed, 7 May 2014 05:50:29 +0000 (14:50 +0900)]
Evas rg_etc1: Also decode RGBA8 from the TGV loader

10 years agoEvas rg_etc2: Complete support for ETC2 with RGBA8_EAC
Jean-Philippe Andre [Wed, 7 May 2014 05:47:45 +0000 (14:47 +0900)]
Evas rg_etc2: Complete support for ETC2 with RGBA8_EAC

Add support for alpha in the ETC2 decoder.

10 years agoEvas rg_etc2: Use new ETC2 decoder for RGB8_ETC2
Jean-Philippe Andre [Wed, 30 Apr 2014 08:18:50 +0000 (17:18 +0900)]
Evas rg_etc2: Use new ETC2 decoder for RGB8_ETC2

Alpha support is still a TODO.

10 years agoEvas rg_etc2: Add ETC2 decoding routines
Jean-Philippe Andre [Wed, 30 Apr 2014 05:10:33 +0000 (14:10 +0900)]
Evas rg_etc2: Add ETC2 decoding routines

Simple implementation of the OpenGL ES 3.0 specification, annex C.1.

10 years agoEvas rg_etc1: Unroll the decoding loop for performance
Matvey Konovalov [Wed, 30 Apr 2014 05:08:55 +0000 (14:08 +0900)]
Evas rg_etc1: Unroll the decoding loop for performance

Measurements have shown a 40% perf increase with these changes.
Patch by Matvey Konovalov.

Signed-off-by: Jean-Philippe Andre <jp.andre@samsung.com>
10 years agoUpdating galician, lithuanian and adding turkish translations
maxerba [Mon, 26 May 2014 19:36:43 +0000 (21:36 +0200)]
Updating galician, lithuanian and adding turkish translations

10 years agoedje should only match swallow visibility if it's actually a group part
Mike Blumenkrantz [Mon, 26 May 2014 16:23:29 +0000 (12:23 -0400)]
edje should only match swallow visibility if it's actually a group part

fixes visibility of actual swallow parts

10 years agoRevert "Revert "recursively hide edje group objects on edje hide""
Mike Blumenkrantz [Mon, 26 May 2014 16:25:07 +0000 (12:25 -0400)]
Revert "Revert "recursively hide edje group objects on edje hide""

This reverts commit b8ca758658cc71bdfecd45bc2a5b8d31ca5e17fc.

10 years agoRevert "recursively hide edje group objects on edje hide"
Carsten Haitzler (Rasterman) [Mon, 26 May 2014 14:17:33 +0000 (23:17 +0900)]
Revert "recursively hide edje group objects on edje hide"

Unless you like your desktop to look something like:

http://www.enlightenment.org/ss/e-53834e0c0cb3c1.87889787.png

then this commit really needs reverting.

This reverts commit 385326b8e4d7c05794670773aad6ece236ea71d9.

10 years agoupdate po's
Carsten Haitzler (Rasterman) [Mon, 26 May 2014 07:41:39 +0000 (16:41 +0900)]
update po's

10 years agorecursively hide edje group objects on edje hide
zmike [Thu, 22 May 2014 17:39:14 +0000 (13:39 -0400)]
recursively hide edje group objects on edje hide

this massively improves edje performance when using groups, which previously would continue calculating their parts even when their parent object was hidden

CPU usage in my test case went from 20-30% to 1%.

@fix

10 years agoEolian: fix generation when optimization is set.
Daniel Zaoui [Mon, 26 May 2014 12:37:33 +0000 (15:37 +0300)]
Eolian: fix generation when optimization is set.

When using -O2 or -O3, the Eina_Bool legacy_support (unsigned char) was
overriding the int eo_needed.

The result was a failure during options check:
Eo flag is not specified (use --eo). Aborting eo generation.

@fix

10 years agoEolian/Generator: add support to namespaces.
Daniel Zaoui [Sun, 25 May 2014 13:44:50 +0000 (16:44 +0300)]
Eolian/Generator: add support to namespaces.

The C generators now support namespaces generation.

10 years agoEolian: add support of namespaces.
Daniel Zaoui [Sun, 25 May 2014 10:53:49 +0000 (13:53 +0300)]
Eolian: add support of namespaces.

It is now possible to define a class in a .eo file as e.g:
Elm::Widgets::Button.
Elm and Widgets will be the namespaces of the class Button.

@feature

10 years agoEolian: remove unused structures.
Daniel Zaoui [Sun, 25 May 2014 08:39:33 +0000 (11:39 +0300)]
Eolian: remove unused structures.

These structs were belonging to a feature that has been removed cause
not considered as needed.

10 years agoEolian/Tests: add test for events.
Daniel Zaoui [Fri, 23 May 2014 14:38:39 +0000 (17:38 +0300)]
Eolian/Tests: add test for events.

10 years agoEolian: modify API to allow future scalability
Daniel Zaoui [Thu, 22 May 2014 14:24:50 +0000 (17:24 +0300)]
Eolian: modify API to allow future scalability

Until now, the functions giving access to class information were taking
the class name as parameter.
Except the fact that we needed to search into a hash table for the internal
class structure, no flexibility is possible.

This change consists in modifying most of the APIs using the class name
with a new Eolian_Class type and adapt the code of the C and C++
generators accordingly.

10 years agoEolian/Tests: add test for overriding functions
Daniel Zaoui [Wed, 21 May 2014 10:56:31 +0000 (13:56 +0300)]
Eolian/Tests: add test for overriding functions

10 years agoEolian/Tests: add test for consts
Daniel Zaoui [Wed, 21 May 2014 08:39:16 +0000 (11:39 +0300)]
Eolian/Tests: add test for consts

This tests checks the different 'const' possibilities in the .eo file
and their good parsing.

10 years agoEolian: add test for generator.
Daniel Zaoui [Fri, 2 May 2014 16:20:10 +0000 (19:20 +0300)]
Eolian: add test for generator.

The test checks the dev code generator.

10 years agoEolian/Generator: add support for implementation source file.
Daniel Zaoui [Thu, 1 May 2014 12:07:26 +0000 (15:07 +0300)]
Eolian/Generator: add support for implementation source file.

By using -gi option, the generator appends the functions that are
present into the given eo file and missing into the developer file
(given via -o option as an in/out file).

@feature

10 years agoEolian/Tests: move parsing tests to a specific file
Daniel Zaoui [Thu, 1 May 2014 12:05:33 +0000 (15:05 +0300)]
Eolian/Tests: move parsing tests to a specific file

10 years agoEolian: add support for typedefs.
Daniel Zaoui [Wed, 30 Apr 2014 09:03:09 +0000 (12:03 +0300)]
Eolian: add support for typedefs.

It includes parser updates, database fill and tests on basic and complex
types.

One can define types in this way:
type Evas_Coord: int; /* Simple type definition */
type List_Objects: Eina_List * <Eo *>;

@feature

10 years agorelease: Switch back to development mode
Stefan Schmidt [Mon, 26 May 2014 10:45:42 +0000 (12:45 +0200)]
release: Switch back to development mode

Open for 1.11 changes now.

10 years agorelease: v1.10.0 v1.10.0
Stefan Schmidt [Mon, 26 May 2014 09:50:38 +0000 (11:50 +0200)]
release: v1.10.0

10 years agoautotools: Fix eolianc-xx header distribution.
Savio Sena [Mon, 26 May 2014 06:47:44 +0000 (15:47 +0900)]
autotools: Fix eolianc-xx header distribution.

Summary:
Sorry, last patch added the headers to the correct relative
location but replaced dist with nodist which is incorrect. nodist is
for generated headers only.

Reviewers: cedric, raster

Reviewed By: raster

CC: felipealmeida, smohanty, cedric
Differential Revision: https://phab.enlightenment.org/D898

10 years agoautotools: fix Eolian-Cxx installation.
Savio Sena [Sat, 24 May 2014 00:27:47 +0000 (02:27 +0200)]
autotools: fix Eolian-Cxx installation.

Summary: grammar/* is now installed to the correct destination.

Reviewers: tasn, cedric

CC: JackDanielZ, felipealmeida, smohanty, cedric
Differential Revision: https://phab.enlightenment.org/D895

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoefl: remove some warnings.
Savio Sena [Sat, 24 May 2014 00:27:11 +0000 (02:27 +0200)]
efl: remove some warnings.

Summary:
Mainly from the examples but also from libunibreak and tests/eet.

I'm not sure if it's really worth to remove warnings from the examples
-- because it adds pedantic-ness to something supposed to be didatic,
but I leave for you guys to judge.

Reviewers: tasn, cedric

CC: felipealmeida, raster, smohanty, cedric
Differential Revision: https://phab.enlightenment.org/D896

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoeina-cxx: remove a warning.
Savio Sena [Sat, 24 May 2014 00:24:42 +0000 (02:24 +0200)]
eina-cxx: remove a warning.

Summary: When assert is disabled 'diff' variable was unused.

Reviewers: tasn, cedric

CC: JackDanielZ, smohanty, felipealmeida, cedric
Differential Revision: https://phab.enlightenment.org/D893

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoautotools: disable automake portability warnings.
Savio Sena [Sat, 24 May 2014 00:23:47 +0000 (02:23 +0200)]
autotools: disable automake portability warnings.

Summary:
Vanish with these:
src/Makefile_Eolian_Helper.am:15: warning: '%'-style pattern rules are a GNU make extension
src/Makefile.am:36:   'src/Makefile_Eolian.am' included from here
src/Makefile_Eolian.am:42:   'src/Makefile_Eolian_Helper.am' included from here
src/Makefile_Eolian_Helper.am:18: warning: '%'-style pattern rules are a GNU make extension
src/Makefile.am:36:   'src/Makefile_Eolian.am' included from here
src/Makefile_Eolian.am:42:   'src/Makefile_Eolian_Helper.am' included from here
src/Makefile_Eolian_Helper.am:21: warning: '%'-style pattern rules are a GNU make extension

Reviewers: tasn, cedric

CC: JackDanielZ, smohanty, felipealmeida, raster, cedric
Differential Revision: https://phab.enlightenment.org/D894

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoUpdating czech translation
maxerba [Fri, 23 May 2014 21:41:05 +0000 (23:41 +0200)]
Updating czech translation

10 years agoEolian: fix class lookup during parsing.
Daniel Zaoui [Wed, 21 May 2014 12:34:08 +0000 (15:34 +0300)]
Eolian: fix class lookup during parsing.

If classes A and B are stored in a same file, when A inherits from B,
and A needs to be generated, the generator, even if B is parsed, will
not search the class in the database but will look for some b.eo.

This patch fixes that issue by checking the existence of the classes
into the database before looking for a corresponding file.

10 years agoevas/text: fix abnormal ellipsis logic when harfbuzz is enabled.
Youngbok Shin [Wed, 21 May 2014 09:39:36 +0000 (10:39 +0100)]
evas/text: fix abnormal ellipsis logic when harfbuzz is enabled.

Summary:
text_props.len indicates visual length from harfbuzz logic when it is available.
But, input pram "idx" is logical length. So, ellipsis works abnormally in some case.
@fix

Test Plan: D870

Reviewers: tasn, woohyun

CC: cedric, herdsman
Differential Revision: https://phab.enlightenment.org/D871

10 years agoevas/text: add text ellipsis test case for ligatures special case.
Youngbok Shin [Wed, 21 May 2014 09:36:27 +0000 (10:36 +0100)]
evas/text: add text ellipsis test case for ligatures special case.

Summary:
Some fonts has combination information for "ff".
When harfbuzz is enabled with the font, evas text ellipsis logic can be broken.

Reviewers: tasn, woohyun, cedric

Reviewed By: tasn

CC: cedric, herdsman
Differential Revision: https://phab.enlightenment.org/D870

Some comment/commit message improvements by TAsn.

10 years agoEo: Add function name to OP desc on Windows
Jean-Philippe Andre [Mon, 19 May 2014 12:04:32 +0000 (21:04 +0900)]
Eo: Add function name to OP desc on Windows

Match function names when the API pointer is out of range.
Reviewed by TAsn and modified according to his comments :)

Differential Revision: https://phab.enlightenment.org/D876

10 years agoWin32: Fix invalid free in eio monitor
Jean-Philippe Andre [Tue, 20 May 2014 09:00:14 +0000 (18:00 +0900)]
Win32: Fix invalid free in eio monitor

You can't just duplicate a string once and free it three times...

10 years agoedje: Edje_Edit - fix wrong behaviour on draggable.event processing while deleting...
Vorobiov Vitalii [Wed, 21 May 2014 04:13:50 +0000 (06:13 +0200)]
edje: Edje_Edit - fix wrong behaviour on draggable.event processing while deleting parts.

Summary:
the draggable.event field (in edje) is pointing on the id of the part it is assigned to.
The main problem is about deleting parts.
After part being deleted, all other parts (that are below the deleted part) have their id shifted,
but event doesnt shift.

For example,
1) draggable.event at some part pointing at part with id=7,
2) some other part is being deleted
3) part with id=7 have new id now (id=6)
4) draggable.event DOESN'T change it's pointer and still points at part with id=7 (which is now has id=6)
5) if there is no parts with id=7 anymore, then it could give you SEGFAULT someday (or not, depending on luck).

This commit contains next changes:
1) after deleting part, draggable.event also shifts (if required)
2) after deleting part that is being pointed with event by someone, it will remove all event points at it
@fix

Reviewers: cedric, raster, seoz

CC: reutskiy.v.v, cedric
Differential Revision: https://phab.enlightenment.org/D878

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoWin32: Add configuration to cross-compile
Shinwoo Kim [Fri, 16 May 2014 02:26:53 +0000 (11:26 +0900)]
Win32: Add configuration to cross-compile

Signed-off-by: Jean-Philippe Andre <jp.andre@samsung.com>
10 years agoWin32: Fix ecore_evas engine search path
Jean-Philippe Andre [Fri, 16 May 2014 07:32:59 +0000 (16:32 +0900)]
Win32: Fix ecore_evas engine search path

Engines are stored in a lib/ folder, while the main DLL files
are in the bin/ folder, so the engine would never be found.

A solution was to add the proper checkme file in the share
folder, but since this is necessary only for Windows, we
can simply use ../lib instead of using the full eina_prefix
detector.

Thanks vtorri for the review.

10 years agoupdate po's
Carsten Haitzler (Rasterman) [Tue, 20 May 2014 04:56:54 +0000 (13:56 +0900)]
update po's

10 years agoautotools: Fix Makefile_Ecore_Audio_Cxx.am.
Savio Sena [Tue, 20 May 2014 03:19:25 +0000 (12:19 +0900)]
autotools: Fix Makefile_Ecore_Audio_Cxx.am.

Summary: This patch fixes T1246.

Reviewers: cedric, smohanty, stefan_schmidt, raster

Reviewed By: raster

CC: felipealmeida, cedric
Maniphest Tasks: T1246

Differential Revision: https://phab.enlightenment.org/D873

10 years agoUpdating translations
maxerba [Mon, 19 May 2014 19:52:22 +0000 (21:52 +0200)]
Updating translations

10 years ago release: v1.10.0 beta3 v1.10.0-beta3
Stefan Schmidt [Mon, 19 May 2014 16:24:17 +0000 (18:24 +0200)]
 release: v1.10.0 beta3

10 years agoevas/render - fix segfault because of null cow access.
ChunEon Park [Mon, 19 May 2014 13:42:36 +0000 (22:42 +0900)]
evas/render - fix segfault because of null cow access.

issue ticket: T1252

@fixed

10 years agoefreet: Fix mistakenly inverted behavior in URI handling
Kai Huuhko [Mon, 19 May 2014 11:15:51 +0000 (14:15 +0300)]
efreet: Fix mistakenly inverted behavior in URI handling

Fixes a mistake in commit 78c9a82a142f619ccd0957957351fe0e573de76c

The authority part would get placed in the struct member "path",
and vice versa.

Lesson learned: Don't rewrite your patch late at night after
arc has messed it up.

10 years agoevas/gl_x11 - replace inlist counting to one api call.
ChunEon Park [Mon, 19 May 2014 09:47:20 +0000 (18:47 +0900)]
evas/gl_x11 - replace inlist counting to one api call.

10 years agoefreet: Handle URIs in a more generic way
Kai Huuhko [Mon, 19 May 2014 06:48:41 +0000 (15:48 +0900)]
efreet: Handle URIs in a more generic way

Summary: Fixes T1219

Reviewers: englebass

CC: cedric
Maniphest Tasks: T1219

Differential Revision: https://phab.enlightenment.org/D865

10 years agoupdate po's
Carsten Haitzler (Rasterman) [Mon, 19 May 2014 06:23:16 +0000 (15:23 +0900)]
update po's

10 years agoEolian/Lexer: fix support of const parameters.
Daniel Zaoui [Sun, 18 May 2014 11:16:58 +0000 (14:16 +0300)]
Eolian/Lexer: fix support of const parameters.

The syntax to define a parameter of a property as const is as following:

set {
   data: const;
}

When no space was present between ':' and "const", a syntax error was
occurring.

10 years agoEo: Make it clearer that a NULL deref can't happen.
Tom Hacohen [Fri, 16 May 2014 13:43:09 +0000 (14:43 +0100)]
Eo: Make it clearer that a NULL deref can't happen.

We check _current_table for NULL, and then populate it (it's a global)
through another function, but we don't really check it's not NULL before
using it, we just assume because of an indirect other variable.

This confused coveritiy, can confuse humans too, and in general risky
(if something changes).

CID 1039419

10 years agoEo callbacks: Fix a possible issue of corrupted callback list.
Tom Hacohen [Mon, 30 Sep 2013 13:16:46 +0000 (14:16 +0100)]
Eo callbacks: Fix a possible issue of corrupted callback list.

This happens when we delete more than one callback at the beginning
of the list (i.e 2 or more) in one go.

10 years agoEo: Don't write to pointer after free.
Tom Hacohen [Mon, 30 Sep 2013 13:15:37 +0000 (14:15 +0100)]
Eo: Don't write to pointer after free.

Fixes coverity: CID 1039898.

10 years agoRevert "Eo: Fix deref after free."
Tom Hacohen [Fri, 16 May 2014 13:30:33 +0000 (14:30 +0100)]
Revert "Eo: Fix deref after free."

Coverity found the issue, but it's an issue we've already fixed in the
past. I don't know how it got lost, but it seems like someone did a bad
merge. Probably when migrating to Eo2.

This reverts commit 831c20464d212fde8d6ee125342e90e5e64efc1e.

10 years agoEo: Fix deref after free.
Tom Hacohen [Fri, 16 May 2014 13:27:39 +0000 (14:27 +0100)]
Eo: Fix deref after free.

In some rare cases it was possible for a pointer to be referenced after
it was already freed. This is now fixed thanks to coverity.

@fix

CID 1039898

10 years agoedje_calc: Document some intended fall through cases in switch-case
Stefan Schmidt [Fri, 16 May 2014 13:00:10 +0000 (15:00 +0200)]
edje_calc: Document some intended fall through cases in switch-case

We share code here for different types. Make a note and also marked them as
false positive in coverity.

CID 1039377, CID 1039376

10 years agoethumb: Memset whole contex in md5 implementation
Stefan Schmidt [Fri, 16 May 2014 12:32:51 +0000 (14:32 +0200)]
ethumb: Memset whole contex in md5 implementation

I don't know what they tried here but if we consider the content
of ctx sensitive we should memset the whole thing.

CID 222382

10 years agoecore_imf: Account for terminating null byte on copy
Stefan Schmidt [Fri, 16 May 2014 11:54:54 +0000 (13:54 +0200)]
ecore_imf: Account for terminating null byte on copy

strlen retruns the count with the terminating null byte but strncpy expects
length with it. Account for that.

10 years agoecore_imf: Avoid potential buffer overrun
Stefan Schmidt [Fri, 16 May 2014 11:52:26 +0000 (13:52 +0200)]
ecore_imf: Avoid potential buffer overrun

We can't just blindly copy the full string into a 128 byte buffer. Adding
some check here to actually stay below the buffer limit or return.

CID 1193233

10 years agofix ecore_thread_global_data_wait to work when no data exists yet
Carsten Haitzler (Rasterman) [Thu, 15 May 2014 12:23:51 +0000 (21:23 +0900)]
fix ecore_thread_global_data_wait to work when no data exists yet

@fix

this is based on feedback from wonil choi, where this didn't work
right. this also has an if check for the hash itself.

10 years agofix behavior of edje entry to match other toolkit entires
Carsten Haitzler (Rasterman) [Thu, 15 May 2014 08:43:00 +0000 (17:43 +0900)]
fix behavior of edje entry to match other toolkit entires

this fixes T1229 in the part about the cursor behavior when you press
updown/left/right when a selection is there. @fix

10 years agoedje_calc: Fix invalid parameters during first calc
Jean-Philippe Andre [Thu, 15 May 2014 03:26:44 +0000 (12:26 +0900)]
edje_calc: Fix invalid parameters during first calc

In particular, ellipsis is -1 by default in Evas, but at this
point (first layout calc), the parameters used for recalc are
incomplete and ellipsis would then be 0 by default (calloc).

As a consequence, Edje will call ellipsis_set(0) enabling
ellipsis even on objects that force "ellipsis: -1".

Solution: set all the parameters before entering text/tb calc.

I believe the other changes are only color and image padding
and should not affect recalc_single.

10 years agofix ecore_con_url using 100% cpu and just let fd handler do work
Carsten Haitzler (Rasterman) [Thu, 15 May 2014 06:48:18 +0000 (15:48 +0900)]
fix ecore_con_url using 100% cpu and just let fd handler do work

this fixes T1251 - @fix

10 years agoevas/wayland_shm: unlink the REAL temporary file
U. Artie Eoff [Wed, 14 May 2014 23:13:33 +0000 (16:13 -0700)]
evas/wayland_shm: unlink the REAL temporary file

Signed-off-by: U. Artie Eoff <ullysses.a.eoff@intel.com>
10 years agoEolian/Generator: fix the class description generation.
Daniel Zaoui [Wed, 14 May 2014 11:26:50 +0000 (14:26 +0300)]
Eolian/Generator: fix the class description generation.

The class description was not used during the generation of the legacy
and Eo headers.

10 years agoEvas filters: Fix NULL dereference (legacy_strdup)
Jean-Philippe Andre [Wed, 14 May 2014 08:47:53 +0000 (17:47 +0900)]
Evas filters: Fix NULL dereference (legacy_strdup)

CID 1210809
CID 1210813

+ remove comment

10 years agoEvas filters: Fix NULL dereferences
Jean-Philippe Andre [Wed, 14 May 2014 08:45:20 +0000 (17:45 +0900)]
Evas filters: Fix NULL dereferences

CID 1210807
CID 1210808

10 years agoEvas filters: Remove logically dead code
Jean-Philippe Andre [Wed, 14 May 2014 08:42:54 +0000 (17:42 +0900)]
Evas filters: Remove logically dead code

CID 1181889

10 years agoedje/Edje_Edit: _edje_edit_real_part_add init field of Edje_Real_part
Vyacheslav Reutskiy [Wed, 14 May 2014 09:11:41 +0000 (11:11 +0200)]
edje/Edje_Edit: _edje_edit_real_part_add init field of Edje_Real_part

Summary:
To correcly display the added pats using the edje edit api, update
init the filds of Edje_Real_Part. Initialize the next filds: type and
typedata. It filds initialized accordingly to Edje_Part_Type value.

@fix

Reviewers: raster, cedric, seoz, Hermet

CC: cedric
Differential Revision: https://phab.enlightenment.org/D854

Signed-off-by: Cedric Bail <cedric.bail@free.fr>
10 years agoedje_calc: Don't add padding twice to text objects
Jean-Philippe Andre [Tue, 13 May 2014 02:22:22 +0000 (11:22 +0900)]
edje_calc: Don't add padding twice to text objects

Text objects declared in Edje will see their padding added twice,
as the Evas_Object_Text itself contains the padding already.

This WILL break some EDC files. It's a bug nonetheless.

Should this be backported?

10 years agoedje_text: Remove no-op pad_get
Jean-Philippe Andre [Tue, 13 May 2014 02:24:43 +0000 (11:24 +0900)]
edje_text: Remove no-op pad_get

This call to pad_get is useless.

10 years agoedje_calc: Update some disabled code so it can compile
Jean-Philippe Andre [Tue, 13 May 2014 02:20:01 +0000 (11:20 +0900)]
edje_calc: Update some disabled code so it can compile

Yup, that's basically a no-op but allows other devs to test
the old behaviour of that function as it's still there:
_edje_part_recalc_single_text

10 years agoadd soem class description to eo timers to test if it gets generated
Carsten Haitzler (Rasterman) [Wed, 14 May 2014 05:16:03 +0000 (14:16 +0900)]
add soem class description to eo timers to test if it gets generated

10 years agode-elipsis edje
Mike Blumenkrantz [Tue, 13 May 2014 20:20:41 +0000 (16:20 -0400)]
de-elipsis edje

the word is "ellipsis". it's bad enough that we have released misspelled api, but we can at least spell things correctly internally.

10 years agoedje_cc now prints helpful warning messages when it compiles broken text parts
Mike Blumenkrantz [Tue, 13 May 2014 20:05:39 +0000 (16:05 -0400)]
edje_cc now prints helpful warning messages when it compiles broken text parts

10 years agoedje_cc should throw errors when inheriting a part from the same part
Mike Blumenkrantz [Tue, 13 May 2014 15:15:02 +0000 (11:15 -0400)]
edje_cc should throw errors when inheriting a part from the same part

10 years agorelease: v1.10.0 beta2 v1.10.0-beta2
Stefan Schmidt [Mon, 12 May 2014 12:49:16 +0000 (14:49 +0200)]
release: v1.10.0 beta2