platform/upstream/efl.git
9 years agoecore-drm: add safety checks to all EAPI functions
Mike Blumenkrantz [Fri, 24 Apr 2015 16:51:34 +0000 (12:51 -0400)]
ecore-drm: add safety checks to all EAPI functions

9 years agoecore-con: try to load libcurl.so.5 before .so.4
Raoul Hecky [Fri, 24 Apr 2015 07:12:32 +0000 (09:12 +0200)]
ecore-con: try to load libcurl.so.5 before .so.4

Curl is using a strange soname bumping where soname version is changed by configure when it detects some differences with off_t.
Explanations here: http://curl.haxx.se/dev/readme-curl_off_t.html

As we are not using curl_off_t in our code, we can simply load libcurl.so.5 if present.

9 years agoedje_util: fix memory leak
Andrii Kroitor [Fri, 24 Apr 2015 07:03:16 +0000 (16:03 +0900)]
edje_util: fix memory leak

Summary:
list returned by evas_object_box_children_get should be freed

@fix

Reviewers: cedric, seoz, raster, reutskiy.v.v, Hermet

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2413

9 years agoEvas GL common: Avoid excessive pipe flushes in image free
Jean-Philippe Andre [Thu, 23 Apr 2015 07:14:22 +0000 (16:14 +0900)]
Evas GL common: Avoid excessive pipe flushes in image free

This should boost performance a little bit when an image is
just unref'ed but not deleted (thus doesn't need a pipe flush).

9 years agoEvas GL masking: Improve performance with scaled images
Jean-Philippe Andre [Thu, 23 Apr 2015 07:12:04 +0000 (16:12 +0900)]
Evas GL masking: Improve performance with scaled images

The GL scaled images is a fast path for masking where
the shader scales masks on the fly.
This optimization actually fixes some issues where the current
texture binding was incorrect.

9 years agoFix ecore_con_eet_client_example example compilation error
Felipe Magno de Almeida [Fri, 24 Apr 2015 01:20:16 +0000 (22:20 -0300)]
Fix ecore_con_eet_client_example example compilation error

Added eet to include flags for Makefile.am for ecore examples

9 years agoFix uninitialized member by coverity CID 1294558
Felipe Magno de Almeida [Fri, 24 Apr 2015 00:24:06 +0000 (21:24 -0300)]
Fix uninitialized member by coverity CID 1294558

Initialize the scope member in the default constructor.

9 years agoecore-drm: refactor _ecore_drm_output_backlight_init() to use fewer gotos
Mike Blumenkrantz [Thu, 23 Apr 2015 17:21:32 +0000 (13:21 -0400)]
ecore-drm: refactor _ecore_drm_output_backlight_init() to use fewer gotos

9 years agoecore-drm: use eeze_udev_find_by_filter() to only return relevant backlight devices
Mike Blumenkrantz [Thu, 23 Apr 2015 17:17:25 +0000 (13:17 -0400)]
ecore-drm: use eeze_udev_find_by_filter() to only return relevant backlight devices

9 years agoecore-drm: restrict output backlight detection to drm device
Mike Blumenkrantz [Thu, 23 Apr 2015 17:11:22 +0000 (13:11 -0400)]
ecore-drm: restrict output backlight detection to drm device

previously this would pull in any available backlight device regardless
of whether it was actually attached to the device or output in any way

9 years agoevas_text: fix default return value.
Jaehwan Kim [Thu, 23 Apr 2015 13:09:08 +0000 (22:09 +0900)]
evas_text: fix default return value.

@fix

9 years agoecore-file: use hash list for inotify monitor tracking
Mike Blumenkrantz [Thu, 23 Apr 2015 12:54:20 +0000 (08:54 -0400)]
ecore-file: use hash list for inotify monitor tracking

inotify reuses watches for multiple instances on the same file, so
keeping a multimap here is necessary to avoid hash collision

fix T2360

9 years agoecore_timer: fix the default return value.
Jaehwan Kim [Wed, 22 Apr 2015 09:01:14 +0000 (18:01 +0900)]
ecore_timer: fix the default return value.

When this function fails to get the interval value, it should return -1.0.
Currently, the value can be integer(-1.0 has an Error).
Maybe it should be fixed.

@fix

9 years agoeina thread_queue: fix typo in doc.
ChunEon Park [Thu, 23 Apr 2015 02:35:35 +0000 (11:35 +0900)]
eina thread_queue: fix typo in doc.

9 years agoecore-drm: fix uninitialized output mode pointers when creating outputs
Mike Blumenkrantz [Thu, 23 Apr 2015 01:11:51 +0000 (21:11 -0400)]
ecore-drm: fix uninitialized output mode pointers when creating outputs

9 years agoecore-drm: fix ecore_drm_devices_get() to return a const value
Mike Blumenkrantz [Wed, 22 Apr 2015 23:34:19 +0000 (19:34 -0400)]
ecore-drm: fix ecore_drm_devices_get() to return a const value

this is the internal device list, we're pretty hosed if someone decides
to free it because there's no const

9 years agoefreet: fix leak in efreet_util_desktop_exec_find()
Mike Blumenkrantz [Wed, 22 Apr 2015 19:32:49 +0000 (15:32 -0400)]
efreet: fix leak in efreet_util_desktop_exec_find()

@fix

9 years agoecore_drm: Don't crash if a device is assigned a bad output
Derek Foreman [Wed, 22 Apr 2015 18:32:16 +0000 (14:32 -0400)]
ecore_drm: Don't crash if a device is assigned a bad output

Summary:
If an input device is assigned via udev to an output that we can't find
we just assign it to the first output instead.

Reviewers: zmike, devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2408

9 years agoeina thread_queue: fix typo in doc.
ChunEon Park [Wed, 22 Apr 2015 14:11:24 +0000 (23:11 +0900)]
eina thread_queue: fix typo in doc.

9 years agoefl: Move test for gl library to evas-gl-drm engine checks and Fix T2167
Chris Michael [Wed, 22 Apr 2015 13:19:07 +0000 (09:19 -0400)]
efl: Move test for gl library to evas-gl-drm engine checks and Fix T2167

Summary: Since we won't need gl library unless we are building
evas-gl-drm engine, move the checks for with_opengl config option to
the gl_drm engine checking macro. Also, since gl_drm is ONLY supported
if we have --with-opengl=es configure option, let's error out with a
message if someone passes anything Other than with_opengl=es.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore: build newly added Ecore_Con examples.
Cedric BAIL [Wed, 22 Apr 2015 12:58:44 +0000 (14:58 +0200)]
ecore: build newly added Ecore_Con examples.

9 years agoecore_con: example program for ecore_con_eet.
Srivardhan Hebbar [Wed, 22 Apr 2015 12:55:33 +0000 (14:55 +0200)]
ecore_con: example program for ecore_con_eet.

Summary:

Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2323

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeio: add testcases for various eio_xattr types set and get functions.
vivek [Wed, 22 Apr 2015 12:52:55 +0000 (14:52 +0200)]
eio: add testcases for various eio_xattr types set and get functions.

Summary:
Added test cases for eio_file_xattr_string_set, eio_file_xattr_string_get, eio_file_xattr_int_set, eio_file_xattr_int_get, eio_file_xattr_double_get, eio_file_xattr_double set functions

Signed-off-by: vivek <vivek.ellur@samsung.com>
Reviewers: cedric

Reviewed By: cedric

Differential Revision: https://phab.enlightenment.org/D2380

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_con: add another test case to cover remaining functions in ecore_con_eet.
Srivardhan Hebbar [Wed, 22 Apr 2015 12:52:28 +0000 (14:52 +0200)]
ecore_con: add another test case to cover remaining functions in ecore_con_eet.

Summary:
When I checked the Jenkins report, I saw few of the functions were left off. So added another case and modified the existing case to cover all the functions or ecore_con_eet.

Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2383

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_con: add test cases to cover more ecore con functions.
Srivardhan Hebbar [Wed, 22 Apr 2015 12:26:57 +0000 (14:26 +0200)]
ecore_con: add test cases to cover more ecore con functions.

Summary:
The dns lookup test would result in a crash unless https://phab.enlightenment.org/D2398 fix.
Also added more functions to existing case to cover ecore_con.c

Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2403

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add test cases for eina file xattr functions.
vivek [Wed, 22 Apr 2015 12:21:36 +0000 (14:21 +0200)]
eina: add test cases for eina file xattr functions.

Summary:
Added test cases for eina_file_xattr_get and eina_file_xattr_value_get functions

Signed-off-by: vivek <vivek.ellur@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2401

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_con: fix dns lookup crash issue.
Srivardhan Hebbar [Wed, 22 Apr 2015 12:19:46 +0000 (14:19 +0200)]
ecore_con: fix dns lookup crash issue.

Summary:
When executed ecore_con_lookup_example function, we were getting a crash. This is the fix for the crash. I analysed it in the valgrind and found the error. Valgrind reported invalid free at this place.
svr->name is getting freed in _ecore_con_server_eo_base_destructor function and svr is getting freed in ecore_con_shutdown function.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2398

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_con: fix ecore_con_lookup_example to actually exit when done.
Srivardhan Hebbar [Wed, 22 Apr 2015 10:41:43 +0000 (12:41 +0200)]
ecore_con: fix ecore_con_lookup_example to actually exit when done.

Summary:
Without ecore_main_loop_quit the example won't end. So added that and as ecore_init is done in ecore_con_init removed that.

Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2399

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add test case for eina_strbuf_string_free function.
vivek [Wed, 22 Apr 2015 10:38:58 +0000 (12:38 +0200)]
eina: add test case for eina_strbuf_string_free function.

Summary:
Added test case for eina_strbuf_string_free function in eina_strbuf module

Signed-off-by: vivek <vivek.ellur@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2382

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: fix rotation and flipping of image in gl engine
kabeer khan [Wed, 22 Apr 2015 09:25:38 +0000 (11:25 +0200)]
evas: fix rotation and flipping of image in gl engine

Summary:
fix rotation(90, 180, 270) and flipping(vertical, horizontal, transpose,
transverse) of evas image in gl engine backend.

@fix

T2338

Signed-off-by: kabeer khan <kabeer.khan@samsung.com>
Reviewers: cedric, jpeg

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2400

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoEldbus: fix potential unsigned int underflow.
Tom Hacohen [Wed, 22 Apr 2015 12:17:28 +0000 (13:17 +0100)]
Eldbus: fix potential unsigned int underflow.

Subtracting two unsigned values results in an unsigned value, which means
abs() will have no effect.
This also means that if base is smaller than size, we'll have an
unsigned int underflow.

I did it this way (with the if) because it looked like base might be
smaller than size. However, please remove the if (and don't reinstate
the abs) if this is not the case.

9 years agoEcore_evas wayland: remove double parenthesis around comparison.
Tom Hacohen [Wed, 22 Apr 2015 12:13:19 +0000 (13:13 +0100)]
Ecore_evas wayland: remove double parenthesis around comparison.

Double praenthesis should be used around assignments, not comparisons.
Doing this correctly all aronud lets the compiler warn us about potential
mistakes.

9 years agoEvas text utils: move the non-ot function into the current ifdef.
Tom Hacohen [Wed, 22 Apr 2015 12:11:52 +0000 (13:11 +0100)]
Evas text utils: move the non-ot function into the current ifdef.

This function should only be build when OT is off. Move
it to the #else part of the ifdef.

9 years agoEina safety: remove unused error string.
Tom Hacohen [Wed, 22 Apr 2015 12:06:16 +0000 (13:06 +0100)]
Eina safety: remove unused error string.

At the moment there are no errors possible for safety checks.
This may change in the future (unlikely), so we have the eina_error
already exposed for that. However, there's no need for the string.

One more down towards a warningless build.

9 years agoEolian generator: remove unusued variable.
Tom Hacohen [Wed, 22 Apr 2015 12:03:17 +0000 (13:03 +0100)]
Eolian generator: remove unusued variable.

9 years agoEdje smart: Add NULL safety
Jean-Philippe Andre [Wed, 22 Apr 2015 01:51:25 +0000 (10:51 +0900)]
Edje smart: Add NULL safety

This case actually happens with E. It's invalid but still happening,
and because Eo doesn't actually check the parent (or its type) fully,
an invalid, non-NULL object was returned.

See T2336 (this is not fixing the root cause!).

9 years agoecore-drm: Change ecore_drm_fb_set warning level
Chris Michael [Tue, 21 Apr 2015 17:58:41 +0000 (13:58 -0400)]
ecore-drm: Change ecore_drm_fb_set warning level

Summary: This is not actually "critical", it just means that we tried
to set a dumb buffer of the wrong size. Ecore_Evas will call this
function again after the canvas has been resized to the proper size,
so just issue a WRN here.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-evas-drm: Fix issue of fullscreen surfaces not getting focus on show
Chris Michael [Tue, 21 Apr 2015 17:42:25 +0000 (13:42 -0400)]
ecore-evas-drm: Fix issue of fullscreen surfaces not getting focus on show

Summary: This fixed T2349 where elm apps would not get initial focus
on show when running via drm. This also addresses the issue of always
sending the state_change event even when not an override.

NB: Thanks to bu5hm4n for reporting ... and even providing a potential fix ;)

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-blend-ops: fix off-by-1 in the C MUL3_SYM "simd" multiplier
Chris Michael [Tue, 21 Apr 2015 15:23:28 +0000 (11:23 -0400)]
evas-blend-ops: fix off-by-1 in the C MUL3_SYM "simd" multiplier

Summary: op_blend functions give off-by-1 using NEON intrinsics in C
MUL3_SYM, this commit fixes that issue with no ill effects on other
platforms.

@fix

Reviewers: raster, cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2394

9 years agoecore_imf: init wayland before using wayland calls.
Marcel Hollerbach [Tue, 21 Apr 2015 14:06:38 +0000 (10:06 -0400)]
ecore_imf: init wayland before using wayland calls.

Summary:
ecore_imf: init wayland if wayland env var exists

Otherwise ecore-wl functions could return some invalid values. The env var check to avoid error messages.

@fix

Reviewers: Sergeant_Whitespace, devilhorns

Reviewed By: Sergeant_Whitespace

Subscribers: Sergeant_Whitespace, cedric

Differential Revision: https://phab.enlightenment.org/D2253

9 years agoEvas 3d: Fix clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 12:08:30 +0000 (21:08 +0900)]
Evas 3d: Fix clang warning

Unused function

9 years agoEdje: Silence more clang stuff
Jean-Philippe Andre [Tue, 21 Apr 2015 11:44:38 +0000 (20:44 +0900)]
Edje: Silence more clang stuff

And make it explicit that the enum is unsigned

9 years agoEvas tiler: Remove unused functions (clang)
Jean-Philippe Andre [Tue, 21 Apr 2015 11:40:41 +0000 (20:40 +0900)]
Evas tiler: Remove unused functions (clang)

rect_list_add_split_strict is still kept for reference in
eina_tiler.c

9 years agoEctor: More clang silencing
Jean-Philippe Andre [Tue, 21 Apr 2015 11:40:27 +0000 (20:40 +0900)]
Ector: More clang silencing

9 years agoEfreet: Silence clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 11:32:54 +0000 (20:32 +0900)]
Efreet: Silence clang warning

[unsigned - unsigned] is unsigned and might wrap around, the implicit
conversion to int (by calling abs()) saved the game here.

9 years agoEeze: Silence clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 11:31:39 +0000 (20:31 +0900)]
Eeze: Silence clang warning

9 years agoEdje edit: Fix clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 11:28:44 +0000 (20:28 +0900)]
Edje edit: Fix clang warning

Remove useless variable

9 years agoEvas: Fix clang warning (actually use function)
Jean-Philippe Andre [Tue, 21 Apr 2015 11:14:23 +0000 (20:14 +0900)]
Evas: Fix clang warning (actually use function)

Those helper functions are barely ever used...

9 years agoEvas GL common: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 10:47:44 +0000 (19:47 +0900)]
Evas GL common: Fix clang warnings

Mooooarrr!

9 years agoEcore_Con: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 11:03:33 +0000 (20:03 +0900)]
Ecore_Con: Fix clang warnings

9 years agoEvas 3d: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 10:45:19 +0000 (19:45 +0900)]
Evas 3d: Fix clang warnings

9 years agoecore_evas_convert: Fix useless clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 10:44:47 +0000 (19:44 +0900)]
ecore_evas_convert: Fix useless clang warning

9 years agoETC1: Fix more clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 10:36:05 +0000 (19:36 +0900)]
ETC1: Fix more clang warnings

Some of these are a bit pointless (eg. init with {0})
but at least this silences clang and helps reveal more
potentially useful warnings.

9 years agoEthumb: Fix clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 10:35:48 +0000 (19:35 +0900)]
Ethumb: Fix clang warning

Unused function.......

9 years agoEvas/cserve2: Fix clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 10:14:12 +0000 (19:14 +0900)]
Evas/cserve2: Fix clang warning

9 years agoEctor: Fix (probably valid) clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 10:02:57 +0000 (19:02 +0900)]
Ector: Fix (probably valid) clang warning

This time it looks like this is a valid warning.
if (abs(double) <= 0.00001f) might too often be true since abs() returns
an int.

9 years agoEctor: Fix clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 09:58:00 +0000 (18:58 +0900)]
Ector: Fix clang warning

Invalid type used for field and implicit conversions

9 years agoETC1/2: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 09:42:33 +0000 (18:42 +0900)]
ETC1/2: Fix clang warnings

Unused functions.
I keep them around for reference, for whoever wants to work on
improving the encoder.

There are still some warnings. Clang is just crazy.

9 years agoEvas DDS: Fix clang warning
Jean-Philippe Andre [Tue, 21 Apr 2015 09:54:27 +0000 (18:54 +0900)]
Evas DDS: Fix clang warning

This one was almost a real catch by clang. Too bad we don't actually
support DXT3.

9 years agoEvas filters: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 09:54:07 +0000 (18:54 +0900)]
Evas filters: Fix clang warnings

9 years agoEvas render2: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 09:44:41 +0000 (18:44 +0900)]
Evas render2: Fix clang warnings

More unused functions. Kept around for later debugging.

9 years agoEolian: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 09:41:48 +0000 (18:41 +0900)]
Eolian: Fix clang warnings

Return proper type (NULL, not bool)

9 years agoEina: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 09:38:50 +0000 (18:38 +0900)]
Eina: Fix clang warnings

Use fabs for double values

9 years agoEina: Fix clang warnings
Jean-Philippe Andre [Tue, 21 Apr 2015 09:35:45 +0000 (18:35 +0900)]
Eina: Fix clang warnings

Remove unused static function

9 years agoEvas: Fix clang warning in generated shaders file
Jean-Philippe Andre [Tue, 21 Apr 2015 09:30:58 +0000 (18:30 +0900)]
Evas: Fix clang warning in generated shaders file

Use const char a[] instead of const char const a[].

9 years agoEdje: Fix compilation warnings with clang
Jean-Philippe Andre [Tue, 21 Apr 2015 09:07:17 +0000 (18:07 +0900)]
Edje: Fix compilation warnings with clang

Fixes T2343.

Thanks for the report.

9 years agoEdje pick: Fix use after free
Jean-Philippe Andre [Mon, 20 Apr 2015 10:35:04 +0000 (19:35 +0900)]
Edje pick: Fix use after free

Fixes CID 1267458

9 years agoEdje: Remove non-sensical NULL check
Jean-Philippe Andre [Mon, 20 Apr 2015 08:57:11 +0000 (17:57 +0900)]
Edje: Remove non-sensical NULL check

Fixes CID 1261450

9 years agoeet - fix emile changes to not leak in eet_cipher/deciper()
Carsten Haitzler (Rasterman) [Tue, 21 Apr 2015 02:01:54 +0000 (11:01 +0900)]
eet - fix emile changes to not leak in eet_cipher/deciper()

coverity spotted a missing binbuf free for the in binbuf created in
CID 1288921 and CID 1288920. this fixes both of these. leak. new.

9 years agoecore_x - axis events - if event struct alloc fails, free axis data
Carsten Haitzler (Rasterman) [Tue, 21 Apr 2015 01:32:40 +0000 (10:32 +0900)]
ecore_x - axis events - if event struct alloc fails, free axis data

this frees the axis data coming in if struct alloc fails - should
basically never happen, but if it does we don't lose axis data as
caller asumes this data will be freed in the event

related to CID 1256115

9 years agoecore_evas extn module - fix coverity complaint
Carsten Haitzler (Rasterman) [Tue, 21 Apr 2015 01:19:28 +0000 (10:19 +0900)]
ecore_evas extn module - fix coverity complaint

coverity is right. CID 1295139 fixed here. tmpstr_free on sometimes
uninitialized local var.

9 years agoecore-evas-extn: Use eina_mkstemp in place of mkstemp
Chris Michael [Mon, 20 Apr 2015 16:46:48 +0000 (12:46 -0400)]
ecore-evas-extn: Use eina_mkstemp in place of mkstemp

Summary: This fixes Coverity CID1039610 'calling mkstemp without
securely setting umask first'. Since we have eina_mkstemp which
already handles creating temporary files in the proper directory,
let's use it.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoemile: Fix resource leak
Chris Michael [Mon, 20 Apr 2015 17:13:55 +0000 (13:13 -0400)]
emile: Fix resource leak

Summary: This fixes Coverity CID1288919 where buffer variable was
being leaked if emile failed to load the image due to corrupt file.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoemile: Fix resource leak
Chris Michael [Mon, 20 Apr 2015 17:06:41 +0000 (13:06 -0400)]
emile: Fix resource leak

Summary: This fixes Coverity CID1288918 where data_start variable was
being leaked if the rectangles did not intersect.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-evas-x: Fix issue of wrong interface function pointer being set
Chris Michael [Mon, 20 Apr 2015 16:56:46 +0000 (12:56 -0400)]
ecore-evas-x: Fix issue of wrong interface function pointer being set

Summary: This fixes Coverity CID1267461 where the pointer to the
interface shape_input_reset function was being assigned multiple
times. It looks like this is just a copy/paste error.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Use #if 0 to comment out unused function
Chris Michael [Mon, 20 Apr 2015 16:15:13 +0000 (12:15 -0400)]
ecore-drm: Use #if 0 to comment out unused function

Summary: It was mentioned on the mailing list to use #if 0 instead of
commenting the block. This does not change any functionality

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoRevert "ecore-drm: Remove 'defined but not used' function"
Chris Michael [Mon, 20 Apr 2015 16:14:27 +0000 (12:14 -0400)]
Revert "ecore-drm: Remove 'defined but not used' function"

This reverts commit d625167e0707cca3648e9933ef65936043372447.

Reverting this to use #if 0 as suggested by q66

9 years agoecore-drm: Remove 'defined but not used' function
Chris Michael [Mon, 20 Apr 2015 15:47:55 +0000 (11:47 -0400)]
ecore-drm: Remove 'defined but not used' function

Summary: As we no longer require the idler to repaint outputs, comment
out the function that was defined for the idler.

NB: I left it commented (and not totally removed) just in case it is
needed for a later (undiscovered) issue

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoEvas 3d mesh: remove non-implemented functions from "implements".
Tom Hacohen [Mon, 20 Apr 2015 14:00:40 +0000 (15:00 +0100)]
Evas 3d mesh: remove non-implemented functions from "implements".

They are not implemented, so they should not be marked as implemented.
Eo already properly handles functions that are not implemented, this is
completely unnecessary.

9 years agoEctor: remove use of #warning directive as TODO.
Tom Hacohen [Mon, 20 Apr 2015 14:00:13 +0000 (15:00 +0100)]
Ector: remove use of #warning directive as TODO.

There's already a fixme in code, and an error output at runtime.
As discussed on the ML, using #warning this way is bad.

9 years agoeio - fix crashes due to stale eio files in tracking list
Carsten Haitzler (Rasterman) [Mon, 20 Apr 2015 12:46:23 +0000 (21:46 +0900)]
eio - fix crashes due to stale eio files in tracking list

so eio_eet didnt free fiels the same as the rest of eio. it thought it
was special. it thought it could just go free() it's objects (even
though they inherit the core Eio_File object type in their structs).
everyone else ended up calling eio_file_free() EXCEPT eio_eet. so the
eio_(long)_file_set() funcs ewnded up adding this eio file to the
tracking list via eio_file_register() but never unregistered because
they really liked to just do their own thing anyway...

BAD CEDRIC! SPANKING TIME

http://33.media.tumblr.com/3422c76c33c3b9b045f623ff73e0bf8d/tumblr_mhvu61N9br1rbavngo1_500.gi

so this unifies all allocation to now use a single allocator (that
also ensures all eio async io objects are zero'd out), and a SINGLE
free path, and then it all works. no more valgrind complaints on e
shutdown/restart whilst eio "things" were used before or ... maybe
still around.

this should fix T2129

so try again "git master" after this commit.

@fix

9 years agorelease: Update NEWS and bump version for 1.14.0-beta2 release v1.14.0-beta2
Stefan Schmidt [Mon, 20 Apr 2015 10:17:35 +0000 (12:17 +0200)]
release: Update NEWS and bump version for 1.14.0-beta2 release

9 years agoemotion gst1 module - remove printf
Carsten Haitzler (Rasterman) [Mon, 20 Apr 2015 03:14:10 +0000 (12:14 +0900)]
emotion gst1 module - remove printf

this debugging printf seems to have been left around - remove it so we
don't get unexpected noise in gst playback

9 years agoemotion - gst1 module - handle long standing bad map/unmap of gst buf
Carsten Haitzler (Rasterman) [Mon, 20 Apr 2015 03:12:27 +0000 (12:12 +0900)]
emotion - gst1 module - handle long standing bad map/unmap of gst buf

this fixes a pending XXX where we didnt keep the buffer mapped as long
as the evas image object pointed to the video pixel data. this fixes
this along with less over-zealous refinf to make things more obvious
(now last_buffer actualyl really matters and if it was mapped and
refed).

@fix

9 years agoEvas: Fix memory leak
Jean-Philippe Andre [Mon, 20 Apr 2015 08:34:15 +0000 (17:34 +0900)]
Evas: Fix memory leak

Fixes CID 1293000

9 years agoEvas tests: Add callback_del order test
Jean-Philippe Andre [Mon, 20 Apr 2015 06:15:34 +0000 (15:15 +0900)]
Evas tests: Add callback_del order test

9 years agoEvas: Remove callbacks in reverse order with callback_del
Jean-Philippe Andre [Mon, 20 Apr 2015 07:31:36 +0000 (16:31 +0900)]
Evas: Remove callbacks in reverse order with callback_del

This function is deprecated, because del_full should be used instead.
Still, the documentation specifies in which order the callbacks should
be deleted.

Fixes T2077 (not really fixes according to the comment but fixes
according to the doc).

9 years agoEvas tests: Add debug logs telling people to ignore ERR
Jean-Philippe Andre [Mon, 20 Apr 2015 07:02:46 +0000 (16:02 +0900)]
Evas tests: Add debug logs telling people to ignore ERR

Some of the ERR messages are expected and it's part of the
test suite to test for invalid cases (parsing stuff).

9 years agoEina tests: fix warning about pointer signedness
Jean-Philippe Andre [Mon, 20 Apr 2015 06:15:10 +0000 (15:15 +0900)]
Eina tests: fix warning about pointer signedness

9 years agoEina tests: suppress warnings over deprecated functions
Jean-Philippe Andre [Mon, 20 Apr 2015 06:14:47 +0000 (15:14 +0900)]
Eina tests: suppress warnings over deprecated functions

9 years agoEvas GL_X11: Fix memory leak
Jean-Philippe Andre [Mon, 20 Apr 2015 02:44:27 +0000 (11:44 +0900)]
Evas GL_X11: Fix memory leak

9 years agobindings/eo: Only install Eo.hh once to avoid failing double install
Stefan Schmidt [Mon, 20 Apr 2015 07:12:34 +0000 (09:12 +0200)]
bindings/eo: Only install Eo.hh once to avoid failing double install

Eo.hh is listed in both dist_installed_eocxxmainheaders_DATA and
dist_installed_eocxxheaders_DATA this leads to a situation where they file
should be installed but already exists and thus the installation fails.

00:32:27.919 /usr/bin/install: cannot create regular file '/srv/build/workspace/changely_efl_configure_xcb/destdir/include/eo-cxx-1/Eo.hh': File exists
00:32:27.923 Makefile:32620: recipe for target 'install-dist_installed_eocxxmainheadersDATA' failed
00:32:27.923 make[4]: *** [install-dist_installed_eocxxmainheadersDATA] Error

Given that it is the only item in dist_installed_eocxxmainheaders_DATA I suspect
it was but into the other rule by accident.

9 years agoevil: remove unused files (evil_p*) and uneeded evil_uuid.c file (managed by mingw...
Vincent Torri [Sun, 19 Apr 2015 06:28:56 +0000 (08:28 +0200)]
evil: remove unused files (evil_p*) and uneeded evil_uuid.c file (managed by mingw-w64)

Just keeping up with latest mingw-w64.

9 years agoeio: forgot to test when xattr tests are disable.
Cedric BAIL [Sun, 19 Apr 2015 07:50:29 +0000 (09:50 +0200)]
eio: forgot to test when xattr tests are disable.

9 years agoeio: this should not be visible outside of this file.
Cedric BAIL [Sat, 18 Apr 2015 16:48:20 +0000 (18:48 +0200)]
eio: this should not be visible outside of this file.

9 years agoeio: add test cases for eio xattr module.
vivek [Sat, 18 Apr 2015 15:40:59 +0000 (17:40 +0200)]
eio: add test cases for eio xattr module.

Summary:
Added test cases for eio_file_xattr, eio_file_xattr_set and eio_file_xattr_get functions

Signed-off-by: vivek <vivek.ellur@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2374

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeio: reorder include to fix build on Windows.
Cedric BAIL [Sat, 18 Apr 2015 15:28:46 +0000 (17:28 +0200)]
eio: reorder include to fix build on Windows.

9 years agoeio: headers are part of the source code and should be handle like they are.
Vincent Torri [Sat, 18 Apr 2015 15:28:16 +0000 (17:28 +0200)]
eio: headers are part of the source code and should be handle like they are.

9 years agoecore_win32: don't fail if OleInitialize() had been called before.
Adrien Nader [Sat, 18 Apr 2015 15:24:53 +0000 (17:24 +0200)]
ecore_win32: don't fail if OleInitialize() had been called before.

9 years agoecore_file: less confusing documentation of ecore_file_download API
kabeer khan [Fri, 17 Apr 2015 10:06:43 +0000 (12:06 +0200)]
ecore_file: less confusing documentation of ecore_file_download API

Summary: Signed-off-by: kabeer khan <kabeer.khan@samsung.com>

Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2373

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>