platform/upstream/efl.git
9 years agoedje_cc: fix segfault when a program attempts to play a non-registered sound
Jean Guyomarc'h [Mon, 12 Oct 2015 19:04:38 +0000 (12:04 -0700)]
edje_cc: fix segfault when a program attempts to play a non-registered sound

Summary:
When an edje program attempted to play a sound that was not registered
(e.g. in a sounds{} block), edje_cc would segfault instead of throwing
an error.

@fix

Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3171

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore-evas-wayland: Trap for subsequent resize events
Chris Michael [Mon, 12 Oct 2015 18:06:48 +0000 (14:06 -0400)]
ecore-evas-wayland: Trap for subsequent resize events

Summary: If we have already resized this ecore_evas to be what we
want, then there is no point in running the below resize code as we
should already be at the requested size. Add a test at the beginning
to see if we have already set these values

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore_x: Optimize _ecore_x_modifiers_get()
Mrunal Sovani [Mon, 12 Oct 2015 16:05:49 +0000 (12:05 -0400)]
ecore_x: Optimize _ecore_x_modifiers_get()

Summary:
_ecore_x_modifiers_get() causes multiple XGetModifierMapping requests to the
server through _ecore_x_key_mask_get() function. _ecore_x_key_mask_get()
requests XModifierKeymap pointer every time it gets executed.
This can be optimized by a single XGetModifierMapping request in the
_ecore_x_modifiers_get() function itself and then passing the XModifierKeymap
pointer thus retrieved to the _ecore_x_key_mask_get() function.

Reviewers: raster, Hermet, tasn, zmike

Subscribers: yashu21985, alok25, sachin.dev, singh.amitesh, cedric

Differential Revision: https://phab.enlightenment.org/D3166

9 years agoevas-drm: remove useless condition
Boram Park [Mon, 12 Oct 2015 15:10:20 +0000 (11:10 -0400)]
evas-drm: remove useless condition

Summary: 'ob->priv.last > ob->priv.last' is always false

@fix

Reviewers: devilhorns, stefan_schmidt

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3159

9 years agoecore-evas-wayland: Fix common window configure callback to not adjust
Chris Michael [Mon, 12 Oct 2015 14:59:33 +0000 (10:59 -0400)]
ecore-evas-wayland: Fix common window configure callback to not adjust
for framespace

Summary: As we have already adjusted for framespace in various code
leading up to a configure callback, don't adjust for it here. This
fixes an issue where xdg surface window geometry would get incorrect
values which were including framespace. The values of the
xdg_surface_set_window_geometry should be Just the geometry of the
visible window.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-evas-wayland: Don't resize ecore_wl window to include frame size
Chris Michael [Mon, 12 Oct 2015 14:58:00 +0000 (10:58 -0400)]
ecore-evas-wayland: Don't resize ecore_wl window to include frame size

Summary: ecore_wl_window_resize adjusts the internal representation of
the window size so this should not include the "framespace" also. This
fixes an issue where xdg surface would be getting an incorrect window
geometry which included framespace.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wl: Don't update window size during ecore_wl_window_resize
Chris Michael [Mon, 12 Oct 2015 14:55:50 +0000 (10:55 -0400)]
ecore-wl: Don't update window size during ecore_wl_window_resize
function

Summary: As we call ecore_wl_window_update_size from various places
when needed, don't call it inside the resize function. This addresses
an issue where sizes for the Ecore_Wl_Window may be out of sync

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeolian generator: cleaner file write
Daniel Kolesa [Mon, 12 Oct 2015 10:45:38 +0000 (11:45 +0100)]
eolian generator: cleaner file write

9 years agoeolian generator: merge legacy and eo header gen entry points
Daniel Kolesa [Mon, 12 Oct 2015 10:35:59 +0000 (11:35 +0100)]
eolian generator: merge legacy and eo header gen entry points

9 years agorelease: Update NEWS and bump version for 1.16.0-beta1 release v1.16.0-beta1
Stefan Schmidt [Mon, 12 Oct 2015 09:46:06 +0000 (11:46 +0200)]
release: Update NEWS and bump version for 1.16.0-beta1 release

9 years agoEvas filters: Fix some build warnings
Jean-Philippe Andre [Mon, 12 Oct 2015 03:03:52 +0000 (12:03 +0900)]
Evas filters: Fix some build warnings

Thanks Tom for spotting one of those

9 years agoEvas filters: Fix some crash and reduce insanity
Jean-Philippe Andre [Thu, 8 Oct 2015 06:01:21 +0000 (15:01 +0900)]
Evas filters: Fix some crash and reduce insanity

In a rare situation the filter would access an invalid buffer.

Solution: Stop messing with buffer references by properly
referencing and releasing them when not needed, rather
than stealing references and hoping for the best. (There were
flags tracking stolen references, but that was still madness)

9 years agoEvas: Add engine function image_ref
Jean-Philippe Andre [Thu, 8 Oct 2015 05:00:39 +0000 (14:00 +0900)]
Evas: Add engine function image_ref

This is to simplify and fix filters. This is an internal function.

9 years agoEvas filter: Fix proxies of filtered images
Jean-Philippe Andre [Thu, 8 Oct 2015 02:19:50 +0000 (11:19 +0900)]
Evas filter: Fix proxies of filtered images

@fix

9 years agoEvas render: Don't render deleted / invisible snapshots
Jean-Philippe Andre [Thu, 8 Oct 2015 01:57:54 +0000 (10:57 +0900)]
Evas render: Don't render deleted / invisible snapshots

Don't render them. They are not visible

9 years agoEvas render: Minor debug code improvements
Jean-Philippe Andre [Thu, 8 Oct 2015 01:56:35 +0000 (10:56 +0900)]
Evas render: Minor debug code improvements

9 years agoEvas snapshot: Fix proxies of snapshot objects
Jean-Philippe Andre [Wed, 7 Oct 2015 11:03:37 +0000 (20:03 +0900)]
Evas snapshot: Fix proxies of snapshot objects

This was broken because the wrong image size was used
in the proxy's fast path (proxy of another image).
Why did snapshot use a specific surface_w,h instead of
reusing the usual cur->image.w,h? (@cedric)

Simplify code.

9 years agoeeze - fix if formatting to be consistent within file
Carsten Haitzler (Rasterman) [Sat, 10 Oct 2015 01:50:44 +0000 (10:50 +0900)]
eeze - fix if formatting to be consistent within file

some if's were single line:
  if (x) blah();

others were
  if (x)
    blah();

make formatting consistent at any rate.

same for EINA_LIST_FREE() blah();

9 years agofix tom's fix by nmot making e segfault EVERY EXIT
Carsten Haitzler (Rasterman) [Sat, 10 Oct 2015 01:45:30 +0000 (10:45 +0900)]
fix tom's fix by nmot making e segfault EVERY EXIT

when fixing something like a leak .. it might be good to fix it to not
CRASH 100% of the time. a leak (minor) is better than a crash by far.

the strings in the list are stringshared.

fixes 896c18045aafb9c677d8d41b4876c47db2b4d5ce

9 years agoEcore_xcb_keymap: Fix memory leak in _ecore_xcb_keymap_finilize
Mrunal Sovani [Fri, 9 Oct 2015 17:17:45 +0000 (13:17 -0400)]
Ecore_xcb_keymap: Fix memory leak in _ecore_xcb_keymap_finilize

Summary:
xcb_get_modifier_mapping_reply_t *reply is obtained from
xcb_get_modifier_mapping_reply and should be freed after use.

Reviewers: raster, Hermet, tasn, zmike

Subscribers: singh.amitesh, yashu21985, alok25, sachin.dev, cedric

Differential Revision: https://phab.enlightenment.org/D3167

@fix

9 years agoeolian-cxx: Fixed possible exception escaping main
Felipe Magno de Almeida [Fri, 9 Oct 2015 16:54:32 +0000 (13:54 -0300)]
eolian-cxx: Fixed possible exception escaping main

Added try/catch around main to show the reason for the generation failure.

@fix CID 1265600

9 years agoeolian-cxx: Fixed assert which was a tautology
Felipe Magno de Almeida [Fri, 9 Oct 2015 16:48:17 +0000 (13:48 -0300)]
eolian-cxx: Fixed assert which was a tautology

Fixed the assert to the correct check

@fix CID 1316018

9 years agoEvas engine wayland_shm: pass the correct enum value.
Tom Hacohen [Fri, 9 Oct 2015 13:22:36 +0000 (14:22 +0100)]
Evas engine wayland_shm: pass the correct enum value.

Following this patch, the correct enum type is now passed to this
function. It's also passing the same value (0), but now from the correct
enum. This doesn't change behaviour, and looks like what was intended.

This fixes the clang warning.

9 years agoecore-wayland: Comment out unused functions
Chris Michael [Fri, 9 Oct 2015 12:55:26 +0000 (08:55 -0400)]
ecore-wayland: Comment out unused functions

Summary: As we do not listen for surface enter/leave events, we do not
need these functions so comment them out.

Thanks to Tom for the report :)

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Comment out unused functions for session listener
Chris Michael [Fri, 9 Oct 2015 12:54:41 +0000 (08:54 -0400)]
ecore-wayland: Comment out unused functions for session listener

Summary: As session listener is currently disabled in Ecore_Wayland,
we don't need these functions defined so comment them out.

Thanks to Tom for the report :)

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoEvas engine software generic: Fix clang warning.
Tom Hacohen [Fri, 9 Oct 2015 11:41:29 +0000 (12:41 +0100)]
Evas engine software generic: Fix clang warning.

Double parenthesis comparison is reserved for evaluation for assignment, not just comparison.
So for example you should do ((x = 2)) and (x == 2), but not ((x == 2)) or (x = 2).

9 years agoEctor renderer: fix return value to be of the right type.
Tom Hacohen [Fri, 9 Oct 2015 11:37:28 +0000 (12:37 +0100)]
Ector renderer: fix return value to be of the right type.

This function expects a pointer as a return value, not a bool.

@fix.

9 years agoEthumb client: Fix shadow issues (warnings).
Tom Hacohen [Fri, 9 Oct 2015 11:35:55 +0000 (12:35 +0100)]
Ethumb client: Fix shadow issues (warnings).

9 years agoEeze sensor udev: Fix memory leak and possible crash.
Tom Hacohen [Fri, 9 Oct 2015 11:32:37 +0000 (12:32 +0100)]
Eeze sensor udev: Fix memory leak and possible crash.

Due to a misplaced semicolon, EINA_LIST_FREE was doing nothing and only
the last part was freed if there was one, and probably a crash if there
wasn't.

This shows again that having statements without {} is dangerous. We need
to force {} and add a lint rule, or at the very least, add a rule to put
the ; in a new line (like clang warns about by default anyway).

(Same as issue found in previous commit)

@fix

9 years agoEdje cc handlers: Fix memory leak and possible crash.
Tom Hacohen [Fri, 9 Oct 2015 11:29:41 +0000 (12:29 +0100)]
Edje cc handlers: Fix memory leak and possible crash.

Due to a misplaced semicolon, EINA_LIST_FREE was doing nothing and only
the last part was freed if there was one, and probably a crash if there
wasn't.

This shows again that having statements without {} is dangerous. We need
to force {} and add a lint rule, or at the very least, add a rule to put
the ; in a new line (like clang warns about by default anyway).

@fix

9 years agoEctor renderer: fix return value to be of the right type.
Tom Hacohen [Fri, 9 Oct 2015 11:27:30 +0000 (12:27 +0100)]
Ector renderer: fix return value to be of the right type.

This function expects a pointer as a return value, not a bool.

@fix.

9 years agoEio model test: initialise pointer (clang warning).
Tom Hacohen [Fri, 9 Oct 2015 11:23:41 +0000 (12:23 +0100)]
Eio model test: initialise pointer (clang warning).

This is probably not an issue because the function should always
return a value, but initialising this variable silence clang's
warning and is not harmful anyway.

9 years agoEo: Fix Eo on Windows.
Tom Hacohen [Fri, 9 Oct 2015 11:06:09 +0000 (12:06 +0100)]
Eo: Fix Eo on Windows.

We use function names instead of function pointers of Windows, because
of dll import/export issues (more in a comment in eo.c). Before this
commit we were comparing the pointers to the strings instead of the
content in some of the places, which caused op desc lookup not to work.
This fixes that.

Thanks to vtorri for his assistance.

@fix

9 years agoevas_canvas3d: fix mesh being added twice added to scene data
Oleksandr Shcherbina [Thu, 8 Oct 2015 10:38:10 +0000 (12:38 +0200)]
evas_canvas3d: fix mesh being added twice added to scene data

Summary: @fix

Reviewers: cedric, Hermet

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3143

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_con: preventing possible NULL pointer dereference.
Srivardhan Hebbar [Thu, 8 Oct 2015 10:26:55 +0000 (12:26 +0200)]
ecore_con: preventing possible NULL pointer dereference.

Summary:
Checking if client is NULL before freeing is not of much use. If it is NULL, it would have crashed before when it was dereferenced in printf. So checking NULL before dereferencing.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3151

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_exe: correcting spelling mistake in comment.
Srivardhan Hebbar [Thu, 8 Oct 2015 10:26:04 +0000 (12:26 +0200)]
ecore_exe: correcting spelling mistake in comment.

Summary:
Corrected the spelling mistake of mechanism.

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

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3152

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecor_xcb: removing useless assignment.
Srivardhan Hebbar [Thu, 8 Oct 2015 10:23:32 +0000 (12:23 +0200)]
ecor_xcb: removing useless assignment.

Summary:
The assignment of NULL will have no effect on the caller. So removed that statment.

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

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3153

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_xcb: initializing the uninitialized variable.
Srivardhan Hebbar [Thu, 8 Oct 2015 10:21:53 +0000 (12:21 +0200)]
ecore_xcb: initializing the uninitialized variable.

Summary:
realwin is uninitialized and is being checked against 0 at 384. It is set at 377 only if win2 == win, else it remains uninitialized. So initializing it to NULL.
Signed-off-by: Srivardhan Hebbar <sri.hebbar@samsung.com>
Reviewers: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3154

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_xcb: handle realloc failure.
Srivardhan Hebbar [Thu, 8 Oct 2015 10:20:46 +0000 (12:20 +0200)]
ecore_xcb: handle realloc failure.

Summary:
If realloc fails, then lst would be NULL, so assigning lst[i] only if realloc is successful, else assigning lst to previous memory location.

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

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3155

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeio_monitor: fix memory leak in eio monitor module
Vivek Ellur [Thu, 8 Oct 2015 10:17:29 +0000 (12:17 +0200)]
eio_monitor: fix memory leak in eio monitor module

Summary:
Free the allocated memory in eio monitor win32 module.

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

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3160

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeio_monitor: fix memory leak issue
Vivek Ellur [Thu, 8 Oct 2015 10:16:23 +0000 (12:16 +0200)]
eio_monitor: fix memory leak issue

Summary:
@fix

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

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3161

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore_fd: add _ecore_main_fdh_poll_del() when fd handler is deleted.
Hosang Kim [Thu, 8 Oct 2015 10:14:00 +0000 (12:14 +0200)]
ecore_fd: add _ecore_main_fdh_poll_del() when fd handler is deleted.

Summary:
When fd handler is deleted by ECORE_CALLBACK_CANCEL, _ecore_main_fdh_poll_del() is not called.
So fd still exists in epoll's event pool.

Reviewers: raster, seoz, woohyun, Hermet, cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3131

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore-evas x - solve internal window size flickrs in e
Carsten Haitzler (Rasterman) [Fri, 9 Oct 2015 03:52:41 +0000 (12:52 +0900)]
ecore-evas x - solve internal window size flickrs in e

so there is an issue that e brings out where configure events get
queued and deferred AND e ends up requesting a new size, but new size
is wrong as its read from an old event (requested size is updated) and
in the end ecore-evas doesnt request the actual new size because
current w/h is "the same" even though it isn't... bah - it's complex
and a self-feeding event issue. just doing the move/resize solves it.

@fix

9 years agoecore-evas-wayland: Set ee->draw_ok appropriately based on visibility
Chris Michael [Thu, 8 Oct 2015 20:04:50 +0000 (16:04 -0400)]
ecore-evas-wayland: Set ee->draw_ok appropriately based on visibility

Summary: A previous code change to the common code which processes
updates is relying on having the ecore_evas->draw_ok flag set in order
to process updates. This change makes sure that ee->draw_ok is in sync
with canvas visibility.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-wayland-shm: Fix silly typo
Chris Michael [Thu, 8 Oct 2015 19:15:00 +0000 (15:15 -0400)]
evas-wayland-shm: Fix silly typo

Summary: OOpppss, this should be eina_rectangle_free

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-evas-wayland-shm: Check for fullscreen and override windows when
Chris Michael [Thu, 8 Oct 2015 17:37:30 +0000 (13:37 -0400)]
ecore-evas-wayland-shm: Check for fullscreen and override windows when
processing updates

Summary: When processing render updates, we should be checking if the
Ecore_Evas "should be visible" property is set.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-wayland-shm: Don't leak rectangles during flush
Chris Michael [Thu, 8 Oct 2015 17:35:07 +0000 (13:35 -0400)]
evas-wayland-shm: Don't leak rectangles during flush

Summary: As we loop the regions of the output buffer and accumulate
changes to post, we should be freeing the rectangles when we are done
with them.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-wayland-shm: Fix to add support for MERGE_MODE Smart
Chris Michael [Wed, 7 Oct 2015 15:00:46 +0000 (11:00 -0400)]
evas-wayland-shm: Fix to add support for MERGE_MODE Smart

Summary: Seems we were never setting any merge_mode for the
wayland-shm engine. This fix implements setting merge_mode

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoEvas filters: Fix crash with async sw rendering
Jean-Philippe Andre [Wed, 7 Oct 2015 09:08:56 +0000 (18:08 +0900)]
Evas filters: Fix crash with async sw rendering

If the filtered object (text or image object) was deleted, its
output image (cached inside the filter data) would be freed
immediately. This could cause crashes in case of async rendering.

@fix

9 years agoeet - be robust about garbage at the end that looks like an idenity
Carsten Haitzler (Rasterman) [Wed, 7 Oct 2015 08:24:20 +0000 (17:24 +0900)]
eet - be robust about garbage at the end that looks like an idenity

if there is an identity signaure at the end, ONLY check it if it looks
like a real one (correct magic number, cert and sig size fields are
sane etc.). this means eet opens dont fail for files that may have
trailing garbage or padding that is not an eet identity signature.

9 years agoethumb - fix client free issue with pending dbus messages
Carsten Haitzler (Rasterman) [Wed, 7 Oct 2015 06:47:04 +0000 (15:47 +0900)]
ethumb - fix client free issue with pending dbus messages

this fixes T2769

@fix

9 years agoeet - fix signature alignment and position
Carsten Haitzler (Rasterman) [Wed, 7 Oct 2015 05:32:37 +0000 (14:32 +0900)]
eet - fix signature alignment and position

9 years agoEvas textblock: Force relayout during proxy render
Jean-Philippe Andre [Tue, 6 Oct 2015 11:18:32 +0000 (20:18 +0900)]
Evas textblock: Force relayout during proxy render

If the textblock object was not visible in the main canvas, but
still needs to be rendered in a proxy surface, then _relayout may
not have been called. This forces generation of paragraphs based on
the current geometry.

This patch is ugly. I know. This is evas render :)

9 years agoEvas render: Fix another bad proxy clipping issue
Jean-Philippe Andre [Tue, 6 Oct 2015 08:36:29 +0000 (17:36 +0900)]
Evas render: Fix another bad proxy clipping issue

This is an ugly hack to fix an issue reported in D3114. I don't
understand how the proposed patch could even fix anything given
the current situation.

Test case:
 - Create edje object with textblock inside
 - Clip out edje object (--> all children become not visible)
 - Take textblock from edje and set it as source of a proxy
 - Mark proxy as source_clip

Result: Nothing visible.

Expected: Proxy should contain the textblock object, since
  source_clip means we ignore the edje object's clipper, and
  only care about the textblock's clipper (entire canvas).

Here's what was happening:
 - During a first pass, textblock is not visible, cur->cache.clip
   is calculated, marked as clip_use=1 with geom 0,0 0x0
 - In a second pass, the proxy is rendered, which needs to draw
   the textblock in a surface. But cache.clip was used and it was
   wrong.

Solution:
 - Ignore cache.clip when rendering inside a proxy. I'm pretty
   sure there are other instances where cache.clip will still
   be a problem.

Problem: textblock never called relayout since it was not
  visible.

Conclusion: cache.clip needs to die. It's a legacy optimization
  that now causes more issues than it fixes.

9 years agoEDC Doc: Some fixes
Jean-Philippe Andre [Fri, 2 Oct 2015 09:14:25 +0000 (18:14 +0900)]
EDC Doc: Some fixes

We need to move this doc to the wiki and complete it :)

9 years agoecore_x: Remove XPrint usage
Derek Foreman [Tue, 6 Oct 2015 17:13:01 +0000 (13:13 -0400)]
ecore_x: Remove XPrint usage

Summary:
Xprint has been deprecated since 2008.

It's recently (August 2015) been removed from debian.

Reviewers: zmike, devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3150

9 years agoecore-evas-gl-drm: Disable async rendering for EGL
Chris Michael [Tue, 6 Oct 2015 13:39:36 +0000 (09:39 -0400)]
ecore-evas-gl-drm: Disable async rendering for EGL

Summary: Disable async rendering for EGL engine as EGL is sync only.
This fixes gl_drm engine to work (in my tests) using ecore_evas
example apps.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeina_quaternion: add since information for new APIs
Stefan Schmidt [Tue, 6 Oct 2015 11:16:18 +0000 (13:16 +0200)]
eina_quaternion: add since information for new APIs

9 years agoeina_matrix: adjust since version for new APIs in 1.16
Stefan Schmidt [Tue, 6 Oct 2015 11:12:47 +0000 (13:12 +0200)]
eina_matrix: adjust since version for new APIs in 1.16

9 years agopo: update po files
Stefan Schmidt [Tue, 6 Oct 2015 10:23:34 +0000 (12:23 +0200)]
po: update po files

9 years agoEvas image: Fix wrong usage of engine data macro
Dongyeon Kim [Tue, 6 Oct 2015 05:29:57 +0000 (14:29 +0900)]
Evas image: Fix wrong usage of engine data macro

ENDT macro should be passed to engine functions instead of
some other function pointer!
@fix

9 years agoevas/gl_x11: Fix indentation and invalid error check in native set/bind/unbind
Dongyeon Kim [Mon, 5 Oct 2015 06:28:53 +0000 (15:28 +0900)]
evas/gl_x11: Fix indentation and invalid error check in native set/bind/unbind

9 years agoevas sw x11 - xcb code has same changes as xlib with optimizations
Carsten Haitzler (Rasterman) [Tue, 6 Oct 2015 05:47:08 +0000 (14:47 +0900)]
evas sw x11 - xcb code has same changes as xlib with optimizations

9 years agoevas engines - fix partial swap to not skip swaps if full mode used
Carsten Haitzler (Rasterman) [Tue, 6 Oct 2015 05:37:10 +0000 (14:37 +0900)]
evas engines - fix partial swap to not skip swaps if full mode used

@fix

9 years agoevas sw generic - printf debug remove
Carsten Haitzler (Rasterman) [Tue, 6 Oct 2015 03:12:19 +0000 (12:12 +0900)]
evas sw generic - printf debug remove

9 years agoeldbus - only do syntax checks if libdbus is new enough
Carsten Haitzler (Rasterman) [Tue, 6 Oct 2015 02:29:16 +0000 (11:29 +0900)]
eldbus - only do syntax checks if libdbus is new enough

9 years agoecore-wayland: Don't use custom animator when in server mode
Chris Michael [Mon, 5 Oct 2015 20:11:52 +0000 (16:11 -0400)]
ecore-wayland: Don't use custom animator when in server mode

Signed-off-by: Chris Michael <cp.michael@samsung.com>
Reviewed-by: @zmike
necessary to avoid breaking the compositor canvas animator

9 years agoecore-evas-wayland-egl: Add support for custom Ecore_Wl animator
Chris Michael [Mon, 5 Oct 2015 17:53:36 +0000 (13:53 -0400)]
ecore-evas-wayland-egl: Add support for custom Ecore_Wl animator

Summary: This adds support for ecore_wl to handle ecore animators and
thus the egl engine can receive frame callbacks now.

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agorelease: Update NEWS and bump version for 1.16.0-alpha1 release v1.16.0-alpha1
Stefan Schmidt [Mon, 5 Oct 2015 13:52:21 +0000 (15:52 +0200)]
release: Update NEWS and bump version for 1.16.0-alpha1 release

9 years agoevas - region updates - go back to tiler as its faster
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 09:41:00 +0000 (18:41 +0900)]
evas - region updates - go back to tiler as its faster

the overhead didnt show up in y tests. do show up with certain
expedite tests. hmmm. last time i messed with region code it was
actually same speed as tiler. bonus was it was fully accurate.

9 years agoevas software x11 - set merge mode explicitly to smart or env var
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 09:36:18 +0000 (18:36 +0900)]
evas software x11 - set merge mode explicitly to smart or env var

this allows merge mode for software-x11 to be set too like for gl.

@feature

9 years agoevas software x11 - increase shm buffer size for modern displays
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 09:33:03 +0000 (18:33 +0900)]
evas software x11 - increase shm buffer size for modern displays

have a bigger shm buffer for larger displays to recycle shm memory more
for sw rendering

@feature

9 years agoevas rectangle - ensure centers of rects that just move/resize dont rend
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 09:30:34 +0000 (18:30 +0900)]
evas rectangle - ensure centers of rects that just move/resize dont rend

calculating updates has been re-rendering centers fo solid rects if
the solid rects moves/resizes with a lot of overlap and objects
underneath changes - their changes show through. this fixes that

@fix

9 years agoecore docs - adjust docs to be explicit about ecore main loops nesting
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 07:04:31 +0000 (16:04 +0900)]
ecore docs - adjust docs to be explicit about ecore main loops nesting

it seems some people think ecore main loops can nest. they can't. be
explicit in docs.

9 years agoecore - detect nested ecore_main_loop_begins from silly code
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 07:00:47 +0000 (16:00 +0900)]
ecore - detect nested ecore_main_loop_begins from silly code

@feature

9 years agoevas - cutouts - use less cpu by skipping non-intersecting cutout objs
Carsten Haitzler (Rasterman) [Mon, 5 Oct 2015 06:51:02 +0000 (15:51 +0900)]
evas - cutouts - use less cpu by skipping non-intersecting cutout objs

@feature

9 years agoector: fix leak of dash info in Cairo backend.
Cedric BAIL [Sun, 4 Oct 2015 15:36:42 +0000 (17:36 +0200)]
ector: fix leak of dash info in Cairo backend.

CID 1316270

9 years agoEcore con: Prevent possible use of uninitialised buffer on the stack.
Tom Hacohen [Sun, 4 Oct 2015 15:27:52 +0000 (16:27 +0100)]
Ecore con: Prevent possible use of uninitialised buffer on the stack.

I'm not actually sure if it's a false, because finding the possible
options is hard. Just to be safe, it's better to set buf to "" in the
else case. I'm doing this instead of initialising the variable so the
compiler/static analyser will be able to warn us if there are other code
paths that should probably set buf, but don't.

CID 1316016

@fix

9 years agoEcore Con: Fix possible timing attacks.
Tom Hacohen [Sun, 4 Oct 2015 15:16:04 +0000 (16:16 +0100)]
Ecore Con: Fix possible timing attacks.

Gist of it: we check, and then there's a window between our check and
the mkdir. We don't really need it anyway, because we just want to mkdir
and if it exists, just go on and do nothing.

CID 1039559
CID 1039558

@fix

9 years agoEcore pipe: clean up handling of already read information.
Tom Hacohen [Sun, 4 Oct 2015 14:34:48 +0000 (15:34 +0100)]
Ecore pipe: clean up handling of already read information.

Coverity was complaining about a possible integer overflow. This isn't
actually possible, but coverity has no way to know that because we were
in fact using a too big of a type. I fixed it to be the right type so
now everything should work.

CID 98384

@fix

9 years agoevas-3d: fix primitives
perepelits.m [Sun, 4 Oct 2015 14:37:17 +0000 (16:37 +0200)]
evas-3d: fix primitives

Summary:
There were warnings after adding primitive in the same frame more then once, the check was added to avoid it

@fix

Reviewers: cedric, raster, Hermet

Subscribers: cedric, artem.popov

Differential Revision: https://phab.enlightenment.org/D3090

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas_box: fix children size and position calculation when padding is used for horizon...
Subodh Kumar [Sun, 4 Oct 2015 13:23:49 +0000 (15:23 +0200)]
evas_box: fix children size and position calculation when padding is used for horizontal flow layout

Summary:
Fix children size and position calculation when padding is used

For each child size calculation padding is adjusted
but box height should include  padding.
Secondly, x and y position of children should not  include
the vertical and horizonatal padding as child size has already
included the given paddings.

@fix

Test Plan:
Please modify test_box.c file in
function test_box_vert2 as follows:

   bx = elm_box_add(win);
   evas_object_size_hint_weight_set(bx, EVAS_HINT_EXPAND, EVAS_HINT_EXPAND);
   elm_win_resize_object_add(win, bx);
   elm_box_padding_set(bx, 50, 50);
   elm_box_layout_set(bx, evas_object_box_layout_flow_horizontal, NULL, NULL);
   evas_object_show(bx);

Now,
1. open elementary_test
2. box
3. Box vert 2 (observe box is broken)
4. Try resizing the window (observe)

Reviewers: raster, cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3049

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeio: make poll backend pass eio_test_monitor.c test suite.
Cedric BAIL [Sun, 4 Oct 2015 08:43:54 +0000 (10:43 +0200)]
eio: make poll backend pass eio_test_monitor.c test suite.

Summary: The current poll backend code uses only filename when reporting changes, while eio tests (and presumably other backends) use absolute path. This makes poll backends behavior more consistent with other backends.

Test Plan: Ran eio test suite with poll backend.

Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3083

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoEdje test: remove unused variable.
Tom Hacohen [Sun, 4 Oct 2015 14:01:08 +0000 (15:01 +0100)]
Edje test: remove unused variable.

9 years agoEo: Remove dead code.
Tom Hacohen [Sun, 4 Oct 2015 13:56:57 +0000 (14:56 +0100)]
Eo: Remove dead code.

This removes code that became dead in commit:
389c6d35f2908018e8b6ae3f8294c02c6b5073a0
The commit doesn't explain why we don't shrink or grow when using mmap,
but this is how it is. No reason to keep old code there.

CID 1240224

@fix

9 years agoeolian: fix out-of-bounds indexing on tokens
Daniel Kolesa [Sun, 4 Oct 2015 13:52:36 +0000 (14:52 +0100)]
eolian: fix out-of-bounds indexing on tokens

Fixes CID 1324818 @fix

9 years agoRevert "eina: change default log level at which we display backtrace."
Carsten Haitzler (Rasterman) [Sun, 4 Oct 2015 08:56:07 +0000 (17:56 +0900)]
Revert "eina: change default log level at which we display backtrace."

This reverts commit 05c18876b3e2dba5dd7caea2517aaf24d92ab614.

i endabled backtraces by DEFAULT because it is otherwise pretty damned
hard to get them from things like enlightenment, or on tizen apps as
running them on the cmdline and having them work is close to
impossible. yes it's noise. FIX THE ERRORS then. this commit is just
like turning off gcc etc. warnings because the code is too noisy in
compilation with them on. it's a bug. fix it. the bt lets you do that
without gdb or re-executing AGAIN with an env var and HOPING to find
the bug the next time around.

not to mention efl programming docs cover this backtrace and say it
is on by default and how to get useful info out of this. this makes
the elf docs a lie by removing what is documented as default.

@fix

9 years agoeldbus - make method calls less crasy since libdbus likes to abort
Carsten Haitzler (Rasterman) [Sun, 4 Oct 2015 08:53:46 +0000 (17:53 +0900)]
eldbus - make method calls less crasy since libdbus likes to abort

changes zmike made to e's systray make e abort due to libdbus seeing
the bus name as invalid. it actually is. it's
/org/ayatana/NotificationItem/steam for example - instead of
or.whatever.blah - thuis e aborts continually if you run steam at all.
eldbus should be more crash-proof thus this adds checks so dbus fun
doesnt cause an app to crash if it happens to call incorrect dbus names.

@fix

9 years agoevas render - async sw - fix context duplication by using proper dup call
Carsten Haitzler (Rasterman) [Sun, 4 Oct 2015 07:50:58 +0000 (16:50 +0900)]
evas render - async sw - fix context duplication by using proper dup call

valgrind pointed this one out. we access freed memory when we dup a
context because the context CONTAINS ptrs to things like rects for
cutouts. we didnt dup these. use the proper context dup call (and
properly ref pixman color image too). this was a random bug/crash
waiting to happen and valgrind caught it. suprising it hasnt turned up
before :/

@fix

9 years agoexamples: fix eina_tmpstr example after revert of strftime
Stefan Schmidt [Sun, 4 Oct 2015 07:19:53 +0000 (09:19 +0200)]
examples: fix eina_tmpstr example after revert of strftime

Was left broken after the eina_tmpstr_strftime revert.

9 years agoRevert "eina_tmpstr: add eina_tmpstr_strftime"
Tom Hacohen [Fri, 2 Oct 2015 08:30:49 +0000 (09:30 +0100)]
Revert "eina_tmpstr: add eina_tmpstr_strftime"

As agreed on the ML, eina_tmpstr_strftime() should be removed.

This reverts commit abaf29cb768375957c9ee0b64d36034c21c618ea.

9 years agoEvas GL: Fix render: force pixel_get if dirty
Jean-Philippe Andre [Fri, 2 Oct 2015 07:51:43 +0000 (16:51 +0900)]
Evas GL: Fix render: force pixel_get if dirty

All examples and docs point to using only the dirty flag in
order to trigger a redraw of an Evas GL surface. The commit
21c43528234 broke this behaviour (for a good reason, but not
related to Evas GL).

This is a compatibility fix.

9 years agoEvas GL: Add new API to get current Evas GL
Jean-Philippe Andre [Fri, 2 Oct 2015 06:15:40 +0000 (15:15 +0900)]
Evas GL: Add new API to get current Evas GL

While this seems to go against the rest of the API (because we
always pass in the Evas GL object), there is no way right now
fully restore a context if there are multiple Evas GL objects.

For instance, an app can use Evas GL from an Elm GLView, and also
use Cairo with another Evas GL at the same time. In that case Cairo
needs to restore the previous Evas GL but the library had no way
of getting the current Evas GL. This is the equivalent of
eglGetCurrentDisplay().

@feature

9 years agoEvas image: Simplify updates if adding whole region at once
Jean-Philippe Andre [Fri, 2 Oct 2015 07:32:05 +0000 (16:32 +0900)]
Evas image: Simplify updates if adding whole region at once

9 years agoeldbus: fix _eldbus_message_iter_arguments_vget
Marcel Hollerbach [Fri, 2 Oct 2015 07:13:48 +0000 (09:13 +0200)]
eldbus: fix _eldbus_message_iter_arguments_vget

Summary:
there is a problem with _eldbus_message_iter_arguments_vget.

Assume: We have a array of type y, there are 3 elements in the array.

Lets say we are iterating with

printf("%c", dbus_message_iter_get_arg_type(iter));

and

dbus_message_iter_next(iter);

throuw the iter.

You will see that this will output yyy. As we are having 3 times v.
While the signature of the message iterator is y.

If you now call eldbus_message_iter_arguments_get(message, "y", &cont)
it will return false, with the errormessage in line 766, reason for this
is that the type of the signature iterator differs from the message
iterator. Because of the upper example.
So all in all: The signature given to the method has not to be equal to
the signature of the message iterator, it has to be equal to the
iterated signature, which cannot be found easily with the eldbus api.

The solution is to only iterate in the message iterator that long until
the signature iterator is at the end, if the type differs in this
region, give a error and return false. Otherwise return true.

The sad thing about this is that it is a behaviour break, but the
behaviour is borked, so its a fix.

I tested this over a week now, without any problem in efl/elm/e.

Reviewers: stefan_schmidt, zmike

Reviewed By: stefan_schmidt

Subscribers: zmike, ceolin, simotek, DaveMDS, cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3074

9 years agoEdje tests: Test inheritance of flag no_render
Jean-Philippe Andre [Fri, 2 Oct 2015 03:21:56 +0000 (12:21 +0900)]
Edje tests: Test inheritance of flag no_render

9 years agoedje_cc: Fix inheritance of "no_render" flag
Jean-Philippe Andre [Fri, 2 Oct 2015 03:17:34 +0000 (12:17 +0900)]
edje_cc: Fix inheritance of "no_render" flag

@fix

9 years agoedje_cc: Add "render" keyword for lazEDC
Jean-Philippe Andre [Fri, 2 Oct 2015 03:13:56 +0000 (12:13 +0900)]
edje_cc: Add "render" keyword for lazEDC

Since @zmike also thought this could be useful, let's add it.

9 years agoector: don't leak cairo context.
Cedric BAIL [Fri, 2 Oct 2015 00:18:54 +0000 (17:18 -0700)]
ector: don't leak cairo context.

9 years agoeio: improve documentation.
Cedric BAIL [Thu, 1 Oct 2015 21:12:50 +0000 (14:12 -0700)]
eio: improve documentation.

This set of documentation update basically make it clearer that Eio
use Eina and do cross linking of function used by Eio to Eina. It
also copy some of the warning coming from Eina documentation into Eio.