platform/upstream/efl.git
10 years agoeina: work around current strange behavior seen in eina_cow.
Cedric Bail [Wed, 4 Dec 2013 06:36:17 +0000 (15:36 +0900)]
eina: work around current strange behavior seen in eina_cow.

This is not perfect, it will just limit the propagation of the problem
for some time. Yes, it does hide it under the carpet, but that's better
than having a crash. Problem seems to be in Eina_Hash, but is really
difficult to reproduce and fix for the moment.

10 years agoeina: add fuzzy testing of the red black tree.
Cedric Bail [Wed, 4 Dec 2013 06:20:17 +0000 (15:20 +0900)]
eina: add fuzzy testing of the red black tree.

10 years agoevas/event - fix the inccorect event hit area.
ChunEon Park [Tue, 3 Dec 2013 16:04:21 +0000 (01:04 +0900)]
evas/event - fix the inccorect event hit area.

cedriccccccccccccccccccccccccccc!!!!!!!!!!!!!!!!!!

10 years agoevas: bugfix in evas_render of not maintaining changed flags on object correctly.
Cedric Bail [Tue, 3 Dec 2013 07:23:05 +0000 (16:23 +0900)]
evas: bugfix in evas_render of not maintaining changed flags on object correctly.

This bug is particularly visible in EFM video preview ( T 539 ). The problem is
that the logic for changed has evolved over time. At the beginning Evas canvas
was flat and could be handle in an array. It was then not using the changed flag
that much. This day, we are living with a tree and we need to propagate the
changed flag to the parent, so that when we walk them we only need to walk the
active objects and don't spend our time on branch that are completely static.

Sadly things did collide here. We remove all object that have been rendered
from the pending_objects array. That does include any smart object that was
processed even if one of the child was not. Once any of the child of that not
processed object is marked changed, it will be propagated up to the first
parent that is changed. As the parent of that one are marked as not changed
when evas_render walk the tree, he is blocked really early in the process and
never get a chance to detect that the child of a not changed object did change
and tada !

The fix is to add all the parent of all the object that are in the pending_objects
array back into the pending_objects list. So they will always be marked as changed.
Another alternative to this logic would have been to change pending_change to
filter out those and keep them around. I choose the first solution as I think it
will be more robust to catch all the parent in all case.

10 years agogitignore: ignore new async test.
Cedric Bail [Tue, 3 Dec 2013 02:48:48 +0000 (11:48 +0900)]
gitignore: ignore new async test.

10 years agoximmodule: check whether pointer is NULL before accessing in xim_preedit_callback_set
Jihoon Kim [Tue, 3 Dec 2013 04:32:15 +0000 (13:32 +0900)]
ximmodule: check whether pointer is NULL before accessing in xim_preedit_callback_set

ecore_imf_context_data_get () can return NULL, so it should be checked before using the result value.

This fixes CID 1135273

10 years agoMerge branch 'withdrawn'
Rafael Antognolli [Mon, 2 Dec 2013 20:05:05 +0000 (18:05 -0200)]
Merge branch 'withdrawn'

Implement withdrawn_set/unset on Wayland engines. It basically calls
ecore_evas_show/hide. Some fixes were done to show/hide on both EGL and
SHM backends too.

This set of patches should fix T155.

10 years agoecore_evas/wayland: Add a function to cleanup the frame callback.
Rafael Antognolli [Mon, 2 Dec 2013 19:39:24 +0000 (17:39 -0200)]
ecore_evas/wayland: Add a function to cleanup the frame callback.

This callback must be manually destroyed and removed on
ecore_evas_hide(), because it won't be delivered anymore after the
surface is destroyed. If the callback still exists, the engine will find
it and avoid doing a new redraw until it is finally called.

Maybe the correct thing to do is to keep this callback in the
Ecore_Wl_Window struct, and have some functions to set/unset it, so it
gets destroyed when the window is hidden. Or when the surface is
destroyed.

10 years agoecore_evas/wayland_egl: Unset the surface on window hide.
Rafael Antognolli [Mon, 2 Dec 2013 19:42:37 +0000 (17:42 -0200)]
ecore_evas/wayland_egl: Unset the surface on window hide.

We must unset it, so the Evas engine knows that it has to do all the
setup to destroy the EGL surfaces associated with it. And in the case of
an ecore_evas_show, recreate everything again.

10 years agoecore_evas/wayland: Implement ecore_evas_withdrawn_set.
Rafael Antognolli [Sat, 30 Nov 2013 12:14:05 +0000 (10:14 -0200)]
ecore_evas/wayland: Implement ecore_evas_withdrawn_set.

Just call ecore_evas_show/hide inside the function.

10 years agoevas/wayland_shm: Don't use a global var to store the sent buffer.
Rafael Antognolli [Sat, 30 Nov 2013 12:09:38 +0000 (10:09 -0200)]
evas/wayland_shm: Don't use a global var to store the sent buffer.

When an Ecore_Evas is hidden, it will destroy the buffer swapper. When
it's shown again, it will try to attach a new buffer, that can be same
buffer. If that global var is still pointing to the old buffer, it can
match to it again and avoid sending a new buffer. So, just put this sent
buffer var in the buffer swapper, and it will get set to NULL when the
swapper is destroyed and created again.

This should fix an intermitent problem of ecore_evas_show() not always
working after an ecore_evas_hide() on the wayland_shm engine.

10 years agoUpdate Korean translation for the efl1.8 release
Seong-ho Cho [Mon, 2 Dec 2013 10:15:21 +0000 (19:15 +0900)]
Update Korean translation for the efl1.8 release

Summary: Translation patch for the efl 1.8. of course, existing one has no problem, but some of sentence structure is enhanced.

Reviewers: seoz

CC: seoz, cedric
Differential Revision: https://phab.enlightenment.org/D360

10 years agoNEON vectorization: added use of COLSAME define in map routine
Carsten Haitzler (Rasterman) [Mon, 2 Dec 2013 07:33:34 +0000 (16:33 +0900)]
NEON vectorization: added use of COLSAME define in map routine

Reviewers: raster

Reviewed By: raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D341

10 years agoRevert "fix smart sma by allowing more recursion"
Daniel Juyung Seo [Mon, 2 Dec 2013 07:02:23 +0000 (16:02 +0900)]
Revert "fix smart sma by allowing more recursion"

This reverts commit 1319733caeb811cd0d53b589b8a11f4229455ed3.

Let's revert this patch and lower the recursion number.
As we released efl 1.8, let's find the root causes of the problem and fix them.
I already told this revert to raster so I am not mean :)

10 years agoAdding the cubic-bezier curve to edje transitions
Otavio Pontes [Mon, 2 Dec 2013 05:59:43 +0000 (14:59 +0900)]
Adding the cubic-bezier curve to edje transitions

Summary: Adding an option to use a cubic-bezier curve in edje transitions.

Reviewers: Sachiel, cedric, raster

Reviewed By: raster

CC: raster
Differential Revision: https://phab.enlightenment.org/D319

10 years agoEldbus: added function for getting bus unique name
Carsten Haitzler (Rasterman) [Mon, 2 Dec 2013 05:57:07 +0000 (14:57 +0900)]
Eldbus: added function for getting bus unique name

Summary:
Hello guys,

We are now working on a accessibility support for elementary (ATSPI2) and we need following function to correctly register application.

Reviewers: cedric, raster, lucasdemarchi

Reviewed By: raster

Differential Revision: https://phab.enlightenment.org/D327

10 years agoedje_edit: Add getters and setters for fixed param
Vyacheslav Reutskiy [Mon, 2 Dec 2013 05:54:28 +0000 (14:54 +0900)]
edje_edit: Add getters and setters for fixed param

Reviewers: cedric, seoz, raster

Reviewed By: raster

CC: cedric
Differential Revision: https://phab.enlightenment.org/D339

10 years agoFUCK - remove logic to silently add -release names to libtool builds
Carsten Haitzler (Rasterman) [Mon, 2 Dec 2013 03:28:12 +0000 (12:28 +0900)]
FUCK - remove logic to silently add -release names to libtool builds

sneaky hidden m4 rule to ADD -release to shared lib names IF profile
!= dev profile. come on! why do that? seriously. this snuck in and was
undetected because i recompiled things against efl and thus things
linked against the new releasename libs. this requires an efl 1.8.1.
argh!

10 years agoecore ecore_x_e: check 0 and -1 for ecore_x_window_prop_card32_get() return value...
Daniel Juyung Seo [Mon, 2 Dec 2013 03:40:56 +0000 (12:40 +0900)]
ecore ecore_x_e: check 0 and -1 for ecore_x_window_prop_card32_get() return value which mean fail.

- ecore_x_window_prop_card32_get() returns:
  -1 if the property was not successfully fetched.
  0 if the property exists but has no elements.

10 years agoNEWS: added EFL 1.9.0 NEWS entry.
Daniel Juyung Seo [Mon, 2 Dec 2013 03:07:18 +0000 (12:07 +0900)]
NEWS: added EFL 1.9.0 NEWS entry.

10 years agoChangeLog: added 1.8 release entry.
Daniel Juyung Seo [Mon, 2 Dec 2013 03:07:04 +0000 (12:07 +0900)]
ChangeLog: added 1.8 release entry.

10 years agofix unchecked return value of stat in ecore_file_mv
Jihoon Kim [Mon, 2 Dec 2013 00:40:48 +0000 (09:40 +0900)]
fix unchecked return value of stat in ecore_file_mv

stat() function may fail and return an error code

This fixes CID 1039697

10 years agoevil: use the right ifdef.
Cedric Bail [Sun, 1 Dec 2013 11:55:54 +0000 (20:55 +0900)]
evil: use the right ifdef.

10 years agov1.8.0 v1.8.0
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 10:04:31 +0000 (19:04 +0900)]
v1.8.0

10 years agoreadme - make requirements list a bullet list
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 05:53:53 +0000 (14:53 +0900)]
readme - make requirements list a bullet list

10 years agoadd some README dependency notes
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 05:51:23 +0000 (14:51 +0900)]
add some README dependency notes

10 years agoeina - fix set-but-unused warning in eina_cow when valgrind off
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 04:52:53 +0000 (13:52 +0900)]
eina - fix set-but-unused warning in eina_cow when valgrind off

10 years agorelease work - switch to release profile
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 04:50:54 +0000 (13:50 +0900)]
release work - switch to release profile

this now maxes out max log level to 3 (same as dev otherwise)

10 years agoevas event freezing is "wrong" for core events - turn it off
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 04:26:04 +0000 (13:26 +0900)]
evas event freezing is "wrong" for core events - turn it off

this turns off event freezing for SOME event types for eveas canvas
events (was was intended to begin with). this fixes T577

10 years agoevas - pre/post render events are for canvas not object
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 03:50:45 +0000 (12:50 +0900)]
evas - pre/post render events are for canvas not object

the eo event naming was wrong for pre/post render event callbacks.
they are for the canvas as a whole, not per object

10 years agofix freetype header includes to be correct where ft 2.5.1 breaks builds
Carsten Haitzler (Rasterman) [Sun, 1 Dec 2013 03:16:28 +0000 (12:16 +0900)]
fix freetype header includes to be correct where ft 2.5.1 breaks builds

this should fix T591

10 years agofix eina_cow issues. Closes T581, T527
Boris Faure [Mon, 25 Nov 2013 22:33:51 +0000 (23:33 +0100)]
fix eina_cow issues. Closes T581, T527

10 years agoecore/ecore_file: NULL-check for input parameter
Oleksandr Shcherbina [Fri, 29 Nov 2013 13:17:18 +0000 (22:17 +0900)]
ecore/ecore_file: NULL-check for input parameter

Add check by NULL for input parameter 'path' in method ecore_file_mkpath

Reviewers: seoz, Hermet

CC: cedric
Differential Revision: https://phab.enlightenment.org/D349

10 years agoeina: formatting
Sebastian Dransfeld [Fri, 29 Nov 2013 21:20:32 +0000 (22:20 +0100)]
eina: formatting

10 years agoefreet: add async menu test
Sebastian Dransfeld [Fri, 29 Nov 2013 21:20:13 +0000 (22:20 +0100)]
efreet: add async menu test

10 years agoefreet: improve locking
Sebastian Dransfeld [Fri, 29 Nov 2013 21:18:51 +0000 (22:18 +0100)]
efreet: improve locking

Do find and ref inside lock, so no one can ref-- after we find desktop

10 years agoecore/wayland: Use touch_focus for touch_based events.
Rafael Antognolli [Fri, 29 Nov 2013 20:03:25 +0000 (18:03 -0200)]
ecore/wayland: Use touch_focus for touch_based events.

Also call the mouse related events with BTN_LEFT instead of 0, otherwise
the correct button won't be assigned to the event (button 1).

10 years agoecore/wayland: Send a mouse up event when the implicit grab starts.
Rafael Antognolli [Fri, 29 Nov 2013 18:04:18 +0000 (16:04 -0200)]
ecore/wayland: Send a mouse up event when the implicit grab starts.

The previous workaround used to send a mouse up event to the
application, using the pointer enter callback to determine when an
implicit grab has finished. This was "simulating" a mouse up event when
the surface move or resize finished.

However, this doesn't work for touch-based move. The pointer enter isn't
emitted, because the wl_touch interface doesn't have the concept of a
pointer.

Changing this code to make it more similar to what Weston's toytoolkit
does, the mouse up event is sent as soon as the move or resize grab
starts. After that, the compositor takes care of the move/resize almost
entirely.

Should fix T468.

10 years agoecore/wayland: Get the touch up event position from the down_info.
Rafael Antognolli [Thu, 28 Nov 2013 19:53:42 +0000 (17:53 -0200)]
ecore/wayland: Get the touch up event position from the down_info.

down_info is a struct that stores some information about the current
pressed touch events. It should be used for that specific touch point,
instead of the generic input info, when sending a mouse_up event.

10 years agoecore/wayland: Remove list of touch points from input structure.
Rafael Antognolli [Fri, 29 Nov 2013 19:14:43 +0000 (17:14 -0200)]
ecore/wayland: Remove list of touch points from input structure.

We already have the "Ecore_Wl_Mouse_Down_Info" list, that contains this
information.

10 years agoeldbus: Fix crash caused when the object of an monitored proxy is deleted
José Roberto de Souza [Fri, 29 Nov 2013 19:21:14 +0000 (17:21 -0200)]
eldbus: Fix crash caused when the object of an monitored proxy is deleted

This happen because proxy was already freed and we try print some information
about the proxy in error message.

This fix: https://phab.enlightenment.org/T543

10 years agoRevert "workaround for T543"
José Roberto de Souza [Fri, 29 Nov 2013 18:39:10 +0000 (16:39 -0200)]
Revert "workaround for T543"

This reverts commit b663b5b8c7232d38ba257f99490cfa22814d7970.

10 years agoreadme update - mention systemd / udev for eeze deps
Carsten Haitzler (Rasterman) [Fri, 29 Nov 2013 15:16:17 +0000 (00:16 +0900)]
readme update - mention systemd / udev for eeze deps

10 years agov1.8.0-beta2
Carsten Haitzler (Rasterman) [Fri, 29 Nov 2013 10:36:41 +0000 (19:36 +0900)]
v1.8.0-beta2

10 years agoFix some typos/grammar issues in new README
Chris Michael [Fri, 29 Nov 2013 10:35:46 +0000 (10:35 +0000)]
Fix some typos/grammar issues in new README

Signed-off-by: Chris Michael <cp.michael@samsung.com>
10 years agoupdate images.mk
Carsten Haitzler (Rasterman) [Fri, 29 Nov 2013 08:36:13 +0000 (17:36 +0900)]
update images.mk

10 years agoreadme - update readme for efl 1.8
Carsten Haitzler (Rasterman) [Fri, 29 Nov 2013 08:31:33 +0000 (17:31 +0900)]
readme - update readme for efl 1.8

lots of extra documentation and information now in the README

10 years agoevas - fix gl egl textrue from pixmap support y invert query extension
Stanislav Vorobiov [Thu, 28 Nov 2013 09:18:16 +0000 (18:18 +0900)]
evas - fix gl egl textrue from pixmap support y invert query extension

Instead of hardcoding yinvert=1 we should query for
EGL_NOK_texture_from_pixmap extension and if it's present
use EGL_Y_INVERTED_NOK value

10 years agoevas/cserve2: Fix interrupt on select in edje_cc
Jean-Philippe Andre [Fri, 29 Nov 2013 02:45:19 +0000 (11:45 +0900)]
evas/cserve2: Fix interrupt on select in edje_cc

Summary:
Block SIGCHLD during select().
This fixes a bug with edje_cc when EVAS_CSERVE2=1: Fixes T464.

select() used to return prematurately with EINTR because the
app received some unexpected signals. In particular SIGCHLD
is received when a child terminates, but this is not a reason
to cancel the image load.

In theory, all blocked signals in pselect() should be pending
until pselect returns, so any SIGCHLD should still trigger
the app's signal handler.

Reviewers: cedric

CC: raster, cedric
Maniphest Tasks: T464

Differential Revision: https://phab.enlightenment.org/D357

10 years agoevas: roundup the step size of an array to fit in one page of memory.
Cedric Bail [Fri, 29 Nov 2013 06:32:27 +0000 (15:32 +0900)]
evas: roundup the step size of an array to fit in one page of memory.

This make an array stay in 4K instead of 20K before. It should reduce a little
bit our peak memory usage without slowing down things to much.

10 years agoeina: fix rounding logic to avoid making the array to short.
Cedric Bail [Fri, 29 Nov 2013 06:30:32 +0000 (15:30 +0900)]
eina: fix rounding logic to avoid making the array to short.

I have no idea how the previous formula was supposed to work at all, but
this one is the same as our alignof code to make sure we do allocate to
the really nearest size and don't do over allocation. Additionnaly it works.

10 years agoecore_evas: fix wrong check in _ecore_evas_buffer_msg_send
Jihoon Kim [Fri, 29 Nov 2013 04:53:47 +0000 (13:53 +0900)]
ecore_evas: fix wrong check in _ecore_evas_buffer_msg_send

'if (!child_ee)' implies that "child_ee" is equal to NULL, but
child_ee is used in the nest of if statement.

This fixes CID 1039431

10 years agoxim immodule: fix dereference before NULL check issue in client_window_set
Jihoon Kim [Fri, 29 Nov 2013 02:35:44 +0000 (11:35 +0900)]
xim immodule: fix dereference before NULL check issue in client_window_set

The variable 'info' can be NULL, therefore it should be checked before accessing.

This fixes CID 1039640

10 years agoibusimmodule: check whether pointer is NULL or not before accessing.
Jihoon Kim [Fri, 29 Nov 2013 02:23:29 +0000 (11:23 +0900)]
ibusimmodule: check whether pointer is NULL or not before accessing.

The pointer variable str can be NULL, therefore we should check whether it is NULL or not
before accessing.

This fixes CID 1132621, 1132622, 1132623

10 years agodoc: add missing images
Jerome Pinot [Thu, 28 Nov 2013 13:15:37 +0000 (22:15 +0900)]
doc: add missing images

The edje-nested.{png,eps} files were missing in the doc tree. I
regenerated them using the sample code from tutorial_edje_nested.html

10 years agoevas - code refactoring.
ChunEon Park [Thu, 28 Nov 2013 12:24:11 +0000 (21:24 +0900)]
evas - code refactoring.

put the frequently used compare prior to the other.

10 years agofix eina_array_remove to actually realloc down in size to remove bloat
Carsten Haitzler (Rasterman) [Thu, 28 Nov 2013 09:00:25 +0000 (18:00 +0900)]
fix eina_array_remove to actually realloc down in size to remove  bloat

eina_array_remove() didnt ever realloc down unless we went to 0
members. this wasn't very good as you'd expect the array to be reduced
in size if enough items were removed. not only that the old code was
stupid and ALWAYS malloc()ed a new array of the exact same size and
copied items over in the most complex way possible, then freed the old
one. this would have added overhead wherever used (evas_render) that
should not have been there.

this is based on the idea in a patch from
Viacheslav Lvov <v.lvov@samsung.com>, but this is a re-do of it
entirely, reducing the codebase massively even compared to the patch
and making it much simpler to read, maintain, actually reduce memory
and cut overhead.

10 years agoecore_ipc: prevent a double free.
Cedric Bail [Thu, 28 Nov 2013 08:17:09 +0000 (17:17 +0900)]
ecore_ipc: prevent a double free.

According to clang static analyzer it is possible to find a path where
buf and svr->buf are pointing to the same array, better be safe than sorry.
Arguably this code could be more readable if it was using Eina_Binbuf.

10 years agoevas: this value is never read after that point, simplify logic.
Cedric Bail [Thu, 28 Nov 2013 08:04:32 +0000 (17:04 +0900)]
evas: this value is never read after that point, simplify logic.

This has been spotted by clang static analyzer.

10 years agoevas: simplify logic for better readability.
Cedric Bail [Thu, 28 Nov 2013 08:00:35 +0000 (17:00 +0900)]
evas: simplify logic for better readability.

This also should solve some complain from clang static analyzer.

10 years agoedje: make sure there is an initial value to avoid garbage out.
Cedric Bail [Thu, 28 Nov 2013 07:55:24 +0000 (16:55 +0900)]
edje: make sure there is an initial value to avoid garbage out.

Detected by clang static analyzer.

10 years agoecore_con: remove unecessary code.
Cedric Bail [Thu, 28 Nov 2013 07:36:21 +0000 (16:36 +0900)]
ecore_con: remove unecessary code.

10 years agoedje: correctly setup PATH and install Edje file as they are needed by examples.
Cedric Bail [Thu, 28 Nov 2013 07:32:29 +0000 (16:32 +0900)]
edje: correctly setup PATH and install Edje file as they are needed by examples.

This will fix T61.

10 years agoevas/line - fix the line position set problem.
ChunEon Park [Thu, 28 Nov 2013 05:03:05 +0000 (14:03 +0900)]
evas/line - fix the line position set problem.

If the line goes to the 0, 0 then it won't be updated. Now it's fixed.

10 years agoecore ecore_x_e: initialize local variable to avoid having wrong value.
Daniel Juyung Seo [Wed, 27 Nov 2013 12:43:06 +0000 (21:43 +0900)]
ecore ecore_x_e: initialize local variable to avoid having wrong value.

Thanks to devilhorns!

10 years agorecursive monitoring and scanning fix for icons and desktop files
Carsten Haitzler (Rasterman) [Wed, 27 Nov 2013 10:45:29 +0000 (19:45 +0900)]
recursive monitoring and scanning fix for icons and desktop files

this fixes T580 ... or SHOULD fix it. there is recursion detection
code now and it properly follows symlinks and dirs. it also properly
updates the file monitor tree for both icons and desktops and it only
monitors dirs, not files (as a dir picks up changes to child data).
tested and it seems not to recurse into self-referencing symlinks
(once it detects the loop) and detects changes nicely in all my tests.

10 years agoecore_con: fix unitialized use of buffer from strncpy.
Cedric Bail [Wed, 27 Nov 2013 05:30:03 +0000 (14:30 +0900)]
ecore_con: fix unitialized use of buffer from strncpy.

Fix CID 1039725.

10 years agoecore_con: fix non initialized use of buf with strncpy.
Cedric Bail [Wed, 27 Nov 2013 05:27:00 +0000 (14:27 +0900)]
ecore_con: fix non initialized use of buf with strncpy.

This fix CID 1039723.

10 years agoevas: add missing define for some target.
Cedric Bail [Wed, 27 Nov 2013 04:45:50 +0000 (13:45 +0900)]
evas: add missing define for some target.

10 years agoRevert "Revert "evas: allow fuzziness on the texture format returned by GL.""
Cedric Bail [Wed, 27 Nov 2013 02:06:56 +0000 (11:06 +0900)]
Revert "Revert "evas: allow fuzziness on the texture format returned by GL.""

This reverts commit 87385b05c3117aa8d46fd4029bfdeadf3444a7b9 that is necessary for desktop target.

10 years agoevas/cserve2: Update image loaders list
Jean-Philippe Andre [Mon, 25 Nov 2013 02:43:43 +0000 (11:43 +0900)]
evas/cserve2: Update image loaders list

evas_image_load.c's list was updated to match the generic
loaders, in 38dd405712f5306fdb4b30.

The list used by cserve should be the same. Actually, there
should be a common function instead of code duplication here.

10 years agoedje_entry: fix out-ouf-bounds issue in _edje_entry_imf_event_preedit_changed_cb
Jihoon Kim [Wed, 27 Nov 2013 00:31:26 +0000 (09:31 +0900)]
edje_entry: fix out-ouf-bounds issue in _edje_entry_imf_event_preedit_changed_cb

The maximum index of tagname is 7, and preedit_type_size is equal to 8.
Checking "attr->preedit_type <= preedit_type_size" implies that the value of "attr->preedit_type" may be up to 8

This patch fixes CID 1039308

10 years agoevas evas_render: removed unnecessary semicolon from a macro.
Daniel Juyung Seo [Tue, 26 Nov 2013 17:02:41 +0000 (02:02 +0900)]
evas evas_render: removed unnecessary semicolon from a macro.

This fixes coverity CID 1132631.

10 years agoecore ibus_imcontext: print the value of cursor_pos only when cursor_pos
Daniel Juyung Seo [Tue, 26 Nov 2013 16:52:21 +0000 (01:52 +0900)]
ecore ibus_imcontext: print the value of cursor_pos only when cursor_pos
is not null.

This fixes coverity CID 1132624.

10 years agoecore scim_imcontext: do not check null for ic. it was already checked
Daniel Juyung Seo [Tue, 26 Nov 2013 16:48:58 +0000 (01:48 +0900)]
ecore scim_imcontext: do not check null for ic. it was already checked
by EINA_SAFETY_ON_NULL_RETURN.

This fixes coverity CID 1132628.

10 years agoemotion emotion_test: free allocated data when the dependent object is
Daniel Juyung Seo [Tue, 26 Nov 2013 16:39:40 +0000 (01:39 +0900)]
emotion emotion_test: free allocated data when the dependent object is
deleted.

This fixes coverity CID 1099709.

10 years agoecore_evas/wayland: Update comments inside rotation code.
Rafael Antognolli [Tue, 26 Nov 2013 16:10:15 +0000 (14:10 -0200)]
ecore_evas/wayland: Update comments inside rotation code.

10 years agoecore_evas/wayland: Fix non-resize rotation.
Rafael Antognolli [Tue, 26 Nov 2013 16:01:35 +0000 (14:01 -0200)]
ecore_evas/wayland: Fix non-resize rotation.

Fix phab T392.

Notice that it should reopen T359, as it wasn't really fixed, but a
rotation with resize was being used when a non-resized rotation was
requested. The cause of the "protruding surfaces" is likely the fact
that Elementary is setting the opaque regions manually, instead of
leaving it to Ecore_Evas. This must be fixed either inside Elementary
itself, or adding the "surface extents" (shadow/non-visible surface
parts) info to Ecore_Evas too.

10 years agoRevert "evas: allow fuzziness on the texture format returned by GL."
Rafael Antognolli [Tue, 26 Nov 2013 13:41:54 +0000 (11:41 -0200)]
Revert "evas: allow fuzziness on the texture format returned by GL."

This reverts commit 2b5b3438e82a9a1bb4086864660942d3c1ddc113.

Breaks gl engines when using GLES + EGL.

10 years agoEo: Fixed eo_manual_free to always return a value.
Tom Hacohen [Tue, 26 Nov 2013 13:08:55 +0000 (13:08 +0000)]
Eo: Fixed eo_manual_free to always return a value.

I wonder how come this wasn't triggered in 64bit and only in 32.

10 years agoEvas tests: Add canvas free with ref tests.
Tom Hacohen [Tue, 26 Nov 2013 12:26:30 +0000 (12:26 +0000)]
Evas tests: Add canvas free with ref tests.

This should check test for the issue fixed in:
bb4a19b5ad6a8fc6720154700aabfa32f4b71244
8bba92fbca7c721a7582b098698404e491f44941

10 years agoEvas object: Set manual_free to false once obj is detached from canvas.
Tom Hacohen [Tue, 26 Nov 2013 12:16:22 +0000 (12:16 +0000)]
Evas object: Set manual_free to false once obj is detached from canvas.

This means unref will really free the object instead of just let it
linger in memory once the object is unrefed.
This is also correct because once detached from the canvas, Evas has
nothing to do with the object and it shouldn't need the manual free
anymore.

10 years agoEo: Make eo_manual_free() return a success flag.
Tom Hacohen [Tue, 26 Nov 2013 12:10:53 +0000 (12:10 +0000)]
Eo: Make eo_manual_free() return a success flag.

eo_manual_free() can fail in some cases, and it is useful for users of
this API to know about it in order to decide what to do.

10 years agoEvas layer: Fix deletion of canvas with still-refed objects.
Tom Hacohen [Tue, 26 Nov 2013 11:46:49 +0000 (11:46 +0000)]
Evas layer: Fix deletion of canvas with still-refed objects.

This fixes the infinite loop while deleting canvases with still
referenced objects.
The canvas should not assume it can get rid of all of the objects. Some
objects might be referenced. I also added a printf to let the user know
about it.

This fixes T565.

10 years agoevas - fix the memory corruption introduced by evas fonts_zero
ChunEon Park [Tue, 26 Nov 2013 11:32:42 +0000 (20:32 +0900)]
evas - fix the memory corruption introduced by evas fonts_zero

10 years agoEo: Add an error message when old parent is wrong.
Tom Hacohen [Tue, 26 Nov 2013 10:10:24 +0000 (10:10 +0000)]
Eo: Add an error message when old parent is wrong.

This is an addition to raster's fix in
55d8156414012e718a7e7adb3f9cd33e19744dc1
(eo - base class protection against NULL old_parent_pd)

10 years agoChangeLog: fixed formatting of Cedric's last changelog.
Daniel Juyung Seo [Tue, 26 Nov 2013 08:10:14 +0000 (17:10 +0900)]
ChangeLog: fixed formatting of Cedric's last changelog.

10 years agoEcore animator: delete animater's internal timer in _end_tick() in all cases.
Daniel Juyung Seo [Tue, 26 Nov 2013 06:58:24 +0000 (15:58 +0900)]
Ecore animator: delete animater's internal timer in _end_tick() in all cases.

When the ecore_animator_source_set() is called with different sources repeatedly, sometimes internal timer is not deleted and this leads animator misbehavior.
Especially when the source is changed from ECORE_ANIMATOR_SOURCE_TIMER to ECORE_ANIMATOR_SOURCE_CUSTOM before the SOURCE_TIMER's internal timer is deleted, this problem occurs.
In this case, even though _end_tick() is called in ecore_animator_source_set(), the SOURCE_TIMER's timer is not deleted because the source is already changed to CUSTOM.
So we should delete the internal timer in _end_tick() in all cases.

10 years agoeo - base class protection against NULL old_parent_pd
Carsten Haitzler (Rasterman) [Tue, 26 Nov 2013 08:01:26 +0000 (17:01 +0900)]
eo - base class protection against NULL old_parent_pd

bug T569 still shows that we have cases where, during e shutdown, we
still get eo_data_scope_get() return NULL for a parent object.
whatever this is, segfaulting is much worse than protecting and
marching on. so protect

10 years agoevas: allow fuzziness on the texture format returned by GL.
Cedric Bail [Tue, 26 Nov 2013 08:00:24 +0000 (17:00 +0900)]
evas: allow fuzziness on the texture format returned by GL.

10 years agoevas - stop over redraw of map content with smart object.
ChunEon Park [Tue, 26 Nov 2013 04:05:05 +0000 (13:05 +0900)]
evas - stop over redraw of map content with smart object.

clip object would always stay in a changed state. nobody was walking them to reset.
Now we do reset there state as soon as we do render them.

10 years agofix clang warning in _edje_entry_focus_in_cb
Jihoon Kim [Mon, 25 Nov 2013 23:48:21 +0000 (08:48 +0900)]
fix clang warning in _edje_entry_focus_in_cb

10 years agoEcore con dns test: Fix test hangs.
Tom Hacohen [Mon, 25 Nov 2013 15:45:01 +0000 (15:45 +0000)]
Ecore con dns test: Fix test hangs.

This should fix the test hangs on Jenkins. Fixed them for me.
The problem was, that the timeout was not handled correctly.
The server object was deleted, but the mainloop was not stopped.

10 years agoEvas textblock: Fixed wrapping of lines ending with whites.
Tom Hacohen [Mon, 25 Nov 2013 14:39:14 +0000 (14:39 +0000)]
Evas textblock: Fixed wrapping of lines ending with whites.

Whites at the end of lines ending with whites should not be cut, but
should be wrapped (there's no legal line break there).

Thanks to Shilpa Singh for reporting.

10 years agoEvas textblock: Added a test for wraps ending with whites.
Tom Hacohen [Mon, 25 Nov 2013 14:38:05 +0000 (14:38 +0000)]
Evas textblock: Added a test for wraps ending with whites.

This tests for word wraps of lines ending with whites.
Those whites should not be cut, they should be wrapped as well.

10 years agoevas - gl engine - async texture upload - protect null tex handle access
Carsten Haitzler (Rasterman) [Mon, 25 Nov 2013 13:29:12 +0000 (22:29 +0900)]
evas - gl engine - async texture upload - protect null tex handle access

this stops a very blantant segv when accessing async->tex when it is
NULL on _evas_gl_preload_main_loop_wakeup().

10 years agoecore_file: Print more informative message when inotify_add_watch was failed.
Ryuan Choi [Mon, 25 Nov 2013 10:21:52 +0000 (19:21 +0900)]
ecore_file: Print more informative message when inotify_add_watch was failed.

inotify_add_watch can be failed because of several reasons.

10 years agoFix keyrepeat going crazy ;)
Chris Michael [Mon, 25 Nov 2013 07:47:52 +0000 (07:47 +0000)]
Fix keyrepeat going crazy ;)

Previously, if you were hold down shift for 1-2 seconds and then press
a key, you would get superfluous key repeats (even tho you released
the printable key). This was because the "key repeat" code was not
checking for the same key before (re)starting the repeat timer.

This fixes the repeating key issue by checking if the key pressed is
different than the one already pressed. If so, it will (re)start the
timer. If it is not different, then the timer is already running and
we don't need to do anything.

Fixes T552

Signed-off-by: Chris Michael <cp.michael@samsung.com>
10 years agoFix ecore wayland key events to fill in Ecore_Event_Key with X11-like
Chris Michael [Mon, 25 Nov 2013 07:35:13 +0000 (07:35 +0000)]
Fix ecore wayland key events to fill in Ecore_Event_Key with X11-like
values.

Previously, the keyname and key fields of the Ecore_Event_Key
structure were being filled in with the capitalized version of the
key. This is due to xkb_keysym_get_name always returning keys with the
modifier applied. There is no actual function in xkbcommon to Not do
this :/ so we have to manually check if Shift is pressed, and if so
then we need to convert the key to lowercase.

Fixes T550

Signed-off-by: Chris Michael <cp.michael@samsung.com>
10 years agoevas - code refactoring.
ChunEon Park [Mon, 25 Nov 2013 07:05:12 +0000 (16:05 +0900)]
evas - code refactoring.

merge duplicated(almost) functions to one.