platform/upstream/efl.git
9 years agoeina - mmap safety handler - fix to only handle sigbus's from io
Carsten Haitzler (Rasterman) [Tue, 13 Jan 2015 10:26:55 +0000 (19:26 +0900)]
eina - mmap safety handler - fix to only handle sigbus's from io

@fix

this fixes the eina mmap safety handler to only  handle what comes
from actual i/o errors (not actual corrupt/hw memory faults or invalid
objects), and if it's not a sigbus eina mmap should handle then pass
it onto the previous handler before eina mmap took over.

9 years agoecore_cocoa: Depend on ecore_cocoa_window.h file for library build.
Stefan Schmidt [Tue, 13 Jan 2015 09:57:53 +0000 (10:57 +0100)]
ecore_cocoa: Depend on ecore_cocoa_window.h file for library build.

Without this missing dependency the file would not show up in the
tarball and the build on MAC OSX would fail.

Thanks a lot to Leif Middelschulte for testing and the report!

9 years agoedje: Remove dead code
Chris Michael [Mon, 12 Jan 2015 16:44:06 +0000 (11:44 -0500)]
edje: Remove dead code

Summary: This removes the commented out dead code block from
9cdb8aaf9bec83a0e0879c04f66408f6326fa2b2

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoEo: use int for _eo_init_count intsead of Eina_Bool
Nicolas Aguirre [Mon, 12 Jan 2015 13:43:25 +0000 (14:43 +0100)]
Eo: use int for _eo_init_count intsead of Eina_Bool

9 years agoEo base class: Fix compliation.
Tom Hacohen [Mon, 12 Jan 2015 13:43:59 +0000 (13:43 +0000)]
Eo base class: Fix compliation.

@inout also used to affect the type generated.
Compile check even the simplest changes.

9 years agoEo base: Remove @inout usage.
Tom Hacohen [Mon, 12 Jan 2015 13:33:39 +0000 (13:33 +0000)]
Eo base: Remove @inout usage.

First step towards deprecation of @inout.

9 years agoconfigure: Makre sure the tarballs are marked as alpha1 v1.13.0-alpha1
Stefan Schmidt [Mon, 12 Jan 2015 12:30:20 +0000 (13:30 +0100)]
configure: Makre sure the tarballs are marked as alpha1

9 years agorelease: Update NEWS and bump version for 1.13.0 release
Stefan Schmidt [Mon, 12 Jan 2015 11:59:42 +0000 (12:59 +0100)]
release: Update NEWS and bump version for 1.13.0 release

9 years agoEo tests: Fix bad free in eo_test_value
Avi Levin [Mon, 12 Jan 2015 11:34:01 +0000 (13:34 +0200)]
Eo tests: Fix bad free in eo_test_value

When running eo_test_suite we get an memory error: "double free or corruption".
That error arises because we try to free an Eina_Value value in eo_test_value
that doesn't need to freed.
I switched the eina_value_free to wina_value_flush, the proper way of
releasing it.

@fix

9 years agoefl - edje_cc - fix default state checks in previous commit
Carsten Haitzler (Rasterman) [Mon, 12 Jan 2015 09:01:34 +0000 (18:01 +0900)]
efl - edje_cc - fix default state checks in previous commit

this fixes 0cb33a46758bd1f66653e97d7ad027a9529b1279
@fix

9 years agoedje - edje_cc - fix wrong state lists where default is not the first
Carsten Haitzler (Rasterman) [Mon, 12 Jan 2015 06:54:22 +0000 (15:54 +0900)]
edje - edje_cc - fix wrong state lists where default is not the first

this fixes T1926
@fix

9 years agoEcore: add EAPI to retrieve Ecore_Evas parent of a buffer.
Daniel Zaoui [Thu, 8 Jan 2015 09:47:31 +0000 (11:47 +0200)]
Ecore: add EAPI to retrieve Ecore_Evas parent of a buffer.

9 years agoEo: add function name to the eo_do pre and post hooks.
Daniel Zaoui [Mon, 27 Oct 2014 06:44:02 +0000 (08:44 +0200)]
Eo: add function name to the eo_do pre and post hooks.

This is useful to determine the Eolian class/function.

9 years agoecore_imf/scim: Fix logically dead code in scim module
Jihoon Kim [Sat, 10 Jan 2015 00:27:10 +0000 (09:27 +0900)]
ecore_imf/scim: Fix logically dead code in scim module

CID 1261449

@fix

9 years agoedje: Fix Coverity CID1261437
Chris Michael [Fri, 9 Jan 2015 20:28:55 +0000 (15:28 -0500)]
edje: Fix Coverity CID1261437

Summary: Coverity reports idential code for different branches here.
Not sure WHY that was done, but I commented out the existing if
statement for posterity.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas: Fix unintentional integer overflow
Chris Michael [Fri, 9 Jan 2015 20:26:32 +0000 (15:26 -0500)]
evas: Fix unintentional integer overflow

Summary: This fixes CID1261436: unintentional integer overflow

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore_evas/wayland_egl: Set alpha of ecore_evas object if parent alpha is set
kabeer khan [Fri, 9 Jan 2015 18:23:55 +0000 (13:23 -0500)]
ecore_evas/wayland_egl: Set alpha of ecore_evas object if parent alpha is set

Summary:
Resolved FIXME ecore_evas_wayland_egl_new_internal to set alpha if parent alpha is set

@fix

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

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1861

9 years agoexamples/elocation: Remove unused config.h include
Stefan Schmidt [Fri, 9 Jan 2015 15:46:51 +0000 (16:46 +0100)]
examples/elocation: Remove unused config.h include

It is not needed here.

9 years agoexamples/evas-3d-blending: Remove unused global variable.
Stefan Schmidt [Fri, 9 Jan 2015 15:44:56 +0000 (16:44 +0100)]
examples/evas-3d-blending: Remove unused global variable.

This got shadowed in one of the functions and is not used at all.

9 years agoexamples/evas-3d-shadows: Avoid shadowing a global variable.
Stefan Schmidt [Fri, 9 Jan 2015 15:37:35 +0000 (16:37 +0100)]
examples/evas-3d-shadows: Avoid shadowing a global variable.

9 years agoexamples/evas-3d-frustum: Remove unused variables.
Stefan Schmidt [Fri, 9 Jan 2015 15:29:50 +0000 (16:29 +0100)]
examples/evas-3d-frustum: Remove unused variables.

9 years agoelocation: Mark elocation as BETA API for the upcoming 1.13 release
Stefan Schmidt [Fri, 9 Jan 2015 15:22:07 +0000 (16:22 +0100)]
elocation: Mark elocation as BETA API for the upcoming 1.13 release

It is not ready as a public API yet. It needs to run on GeoClue1
and GeoClue2 which will definitely bring API changes as some parts
are tied to mcuh to what GeoClue1 offers right now.

9 years agoelua: convert some eolian binding API results to numbers
Daniel Kolesa [Fri, 9 Jan 2015 15:45:03 +0000 (15:45 +0000)]
elua: convert some eolian binding API results to numbers

This might fix some bugs and is the right thing to do anyway.

9 years agoecore-evas-wayland: Port fix of min/max size calculations to the
Chris Michael [Thu, 8 Jan 2015 14:10:21 +0000 (09:10 -0500)]
ecore-evas-wayland: Port fix of min/max size calculations to the
configure event handler

Summary: This patch ports the fix for windows without a min/max size
being set over to the configure event handler (which was also not
taking into account the fact that Some windows may Not have a min/max
property set on them.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoedje: update the edc reference
Vyacheslav Reutskiy [Fri, 9 Jan 2015 07:42:19 +0000 (07:42 +0000)]
edje: update the edc reference

Add a list of global signals that edje provide.

9 years agotests/ecore_thread: Disable verbose logging with printf
Stefan Schmidt [Thu, 8 Jan 2015 16:03:55 +0000 (17:03 +0100)]
tests/ecore_thread: Disable verbose logging with printf

If we are going to test for 10000000 messages we should not print out for each
of them to stdout. The log file for ecore_suite will catch them all. It ended
up being 199MB on my machine just filled with these debug statements.

I left it in in case someone needs it for debugging but we should not have it
on by default.

9 years agoeo: Fix bad addressing in _eo_classes array
Avi Levin [Thu, 8 Jan 2015 12:40:56 +0000 (14:40 +0200)]
eo: Fix bad addressing in _eo_classes array

The header.id was masked before using it as index in the _eo_classes
array and was not unmasked when used.
It hasn't caused segfault (by sheer luck) but was wrong.

@fix

9 years agoecore_evas_x: Free XCounter values leak.
Myoungwoon Kim [Thu, 8 Jan 2015 13:35:49 +0000 (08:35 -0500)]
ecore_evas_x: Free XCounter values leak.

Summary:
Whenever creating a window via elm_win_add(), save netwm_sync_counter values to edata->netwm_sync_counter
by ecore_x_sync_counter_new().
However, never free the values when a window has been destroyed.
So this patch is for free the values in _ecore_evas_x_free().

@fix

Reviewers: cedric, raster, Hermet, devilhorns

Reviewed By: devilhorns

Subscribers: cedric, seoz

Differential Revision: https://phab.enlightenment.org/D1855

9 years agoecore_imf: Add null check condition in ecore_imf module
vivek [Thu, 8 Jan 2015 13:34:01 +0000 (08:34 -0500)]
ecore_imf: Add null check condition in ecore_imf module

Summary: Fixed null pointer dereference issue in ecore_imf module

@fix

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

Reviewed By: devilhorns

Subscribers: raster, cedric

Differential Revision: https://phab.enlightenment.org/D1813

9 years agoeina: all Eina_Stringshare printf variant now return empty string as expected.
Cedric BAIL [Thu, 8 Jan 2015 11:34:20 +0000 (12:34 +0100)]
eina: all Eina_Stringshare printf variant now return empty string as expected.

Previously Eina_Stringshare printf variant where returning NULL with ("%s", ""). This
is highly disturbing as you would have gotten "(null)", if you asked for ("s", NULL).
To clarify and make things logical, we now return "".

9 years agoeina: fix memory leaks in eina_stringshare_printf/vprintf functions.
Amitesh Singh [Thu, 8 Jan 2015 10:48:01 +0000 (19:48 +0900)]
eina: fix memory leaks in eina_stringshare_printf/vprintf functions.

Summary: @Fix

Test Plan:
  CK_FORK=no valgrind --leak-check=full {UNIT_TEST_PROGRAM} "String Share"

Reviewers: cedric, seoz, Hermet, raster

Subscribers: cedric, seoz

Differential Revision: https://phab.enlightenment.org/D1857

9 years agoevas:fix markup text length issue
Jae Yong Hwang [Thu, 8 Jan 2015 10:38:48 +0000 (19:38 +0900)]
evas:fix markup text length issue

Summary:
There was no conversion to the double quotation mark in the evas_textblock_text_utf8_to_markup function.
The price of the text coming out to API and text coming out to Textblock was different as a result.
As a result, Two text lengths came out differently.
So, I added the exceptional treatment part in the evas_textblock_text_utf8_to_markup function.

@fix

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1853

9 years ago[Evas, Edje] Adding of new structures
perepelits.m [Thu, 8 Jan 2015 06:21:15 +0000 (15:21 +0900)]
[Evas, Edje] Adding of new structures

Summary:
There will be several methods to set orientation in edc, so we have decided to make one big vector,
the main reason is that we use quaternion by default, but look_at, for example, is given as 6 coordinates.

Reviewers: Hermet, cedric, raster

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1693

9 years agoecore: fix compilation, missing return value in ecore_thread_wait
Jérémy Zurcher [Wed, 7 Jan 2015 21:12:29 +0000 (22:12 +0100)]
ecore: fix compilation, missing return value in ecore_thread_wait

@fix ecore_thread_wait return EINA_TRUE if execution is over or thread is NULL

9 years agoecore-evas-wayland: Fix min/max size calculations
Chris Michael [Wed, 7 Jan 2015 19:47:55 +0000 (14:47 -0500)]
ecore-evas-wayland: Fix min/max size calculations

Summary: This fixes an issue for windows which do not set a min or max
size in the properties. This was discovered when running Enlightenment
in a Wayland-Only scenario, and trying to bring up the settings panel
which would cause an endless loop in calculating the proper window
size due to min/max not being set.

@fix

NB: Thanks to Mike for the help in tracing this ! :)

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Simplify opaque and input region handling.
Chris Michael [Wed, 7 Jan 2015 19:46:14 +0000 (14:46 -0500)]
ecore-wayland: Simplify opaque and input region handling.

Summary: We can make opaque and input region handling simplier if we
just use one opaque & input region per window. Previous code would
always create a new region, set it to the surface, then destroy it.
This code works much nicer in that it hammers the protocol with less
region create/destroy calls.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-evas-wayland: Fix formatting
Chris Michael [Wed, 7 Jan 2015 18:37:55 +0000 (13:37 -0500)]
ecore-evas-wayland: Fix formatting

Summary: no functional changes, just formatting

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeio: forcefully wait on shutdown for all thread to stop.
Cedric BAIL [Wed, 7 Jan 2015 15:43:47 +0000 (16:43 +0100)]
eio: forcefully wait on shutdown for all thread to stop.

This should avoid potential crash during shutdown while some Eio thread
were still running. We are still not blocking for more than 30s, so if
an IO is blocked on a dead device, you should be fine.

9 years agoecore: add ecore_thread_wait and necessary infrastructure.
Cedric BAIL [Wed, 7 Jan 2015 15:42:24 +0000 (16:42 +0100)]
ecore: add ecore_thread_wait and necessary infrastructure.

This enable the possibility to block the main loop until a
specific thread is done. It may trigger still process ending
of other thread during that function call, but not any other
type of event (timer, animator, idler, ... are all ignored).

9 years agoecore-evas-drm: Fix formatting
Chris Michael [Wed, 7 Jan 2015 14:52:34 +0000 (09:52 -0500)]
ecore-evas-drm: Fix formatting

Summary: No functional changes, just formatting cleanup

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore_evas_drm: Added support for initial rotation in ecore_evas_drm
kabeer khan [Wed, 7 Jan 2015 14:51:12 +0000 (09:51 -0500)]
ecore_evas_drm: Added support for initial rotation in ecore_evas_drm

Summary:
Resolved FIXME added support for initial rotation in ecore_evas_drm initialisation

@fix

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

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1837

9 years agoevas textgrid - fix drawing of codepoints that are 0
Carsten Haitzler (Rasterman) [Wed, 7 Jan 2015 13:29:00 +0000 (22:29 +0900)]
evas textgrid - fix drawing of codepoints that are 0

this fixes artifacts if you have cells with 0 codepoints and also have
underlines or strikethroughs in them.

@fix

9 years agoevas map aa - match previous fix.
Carsten Haitzler (Rasterman) [Wed, 7 Jan 2015 12:04:14 +0000 (21:04 +0900)]
evas map aa - match previous fix.

this matches 83eb1aa3c2e15568c89546048d72d7cae071ee42 fix in the other
map render func

9 years agoEvas map: Fix AA with opaque images
Jean-Philippe Andre [Wed, 7 Jan 2015 11:21:16 +0000 (20:21 +0900)]
Evas map: Fix AA with opaque images

Well... actually this is not exactly a fix.
It just restores the previous behaviour, and allows AA to
work. As in, it won't draw ugly black lines but properly
blend to transparent.

But there is still a problem:

The image map render function changes the alpha flag on the source
image if AA is enabled or if the map has an alpha color. This is
actually wrong as images forcefully set to not have any alpha
(with evas_object_image_alpha_set(0)) will then not be opaque
anymore.

Right now I can't think of a solution (also I don't quite follow
the entire pipeline in evas map...). Changing the flag will
make some opaque areas transparent. Not changing the flag will
produce ugly artifacts where AA blending should happen. Fix one
bug and the other appears, and vice versa.

This can be tested with the example evas-map-aa and adding an
alpha channel to cube1.png (with gimp for instance) but manually
setting alpha to 0 in the code. Weird stuff will happen (try
playing with the map and pressing I to switch to/from image mode).

9 years agoEvas examples: Fix comment also in evas-map-aa.c
Jean-Philippe Andre [Wed, 7 Jan 2015 08:53:24 +0000 (17:53 +0900)]
Evas examples: Fix comment also in evas-map-aa.c

i'm dumb and didn't even read the line above the one i changed...

9 years agoEvas examples: Fix command line in evas-map-aa.c
Jean-Philippe Andre [Wed, 7 Jan 2015 08:49:02 +0000 (17:49 +0900)]
Evas examples: Fix command line in evas-map-aa.c

9 years agoEvas map: Kill build warnings and don't set image flag
Jean-Philippe Andre [Wed, 7 Jan 2015 08:30:46 +0000 (17:30 +0900)]
Evas map: Kill build warnings and don't set image flag

Don't change a source image alpha flag at render time.
It doesn't make sense at this point.

Also remove unused int pa.

9 years agoevas map - new aa map has extra overflow with line list - fix it.
Carsten Haitzler (Rasterman) [Wed, 7 Jan 2015 08:21:46 +0000 (17:21 +0900)]
evas map - new aa map has extra overflow with line list - fix it.

this fixes a feature added in this version of efl. as above.

9 years agoedje - fix uninitialized coord sizes if swallow obj is invalid
Carsten Haitzler (Rasterman) [Wed, 7 Jan 2015 07:00:59 +0000 (16:00 +0900)]
edje - fix uninitialized coord sizes if swallow obj is invalid

this fixes invalid sizing calcs if shallowed obj is invalid and we
cant get geometry etc.

@fix

9 years agoEvas masking: Fix a potential issue with pixman
Jean-Philippe Andre [Wed, 7 Jan 2015 07:48:39 +0000 (16:48 +0900)]
Evas masking: Fix a potential issue with pixman

Untested. Just skip pixman path when there's a mask.

9 years agoEvas masking: Add some test cases
Jean-Philippe Andre [Tue, 6 Jan 2015 07:58:33 +0000 (16:58 +0900)]
Evas masking: Add some test cases

Here are only 3 very basic test cases.

One is a dumb set/get to check that image objects can
be passed as clippers.

The other one is a pixel verification test with extremely
basic data (NEAREST scaling and just rectangles). It also
compares text clipping and masking.

The last one performs a very basic verification that masks
of masks work.

9 years agoEvas masking: Minor fixes in GL masking
Jean-Philippe Andre [Mon, 5 Jan 2015 04:44:32 +0000 (13:44 +0900)]
Evas masking: Minor fixes in GL masking

9 years agoEvas masking: Fix invalid geometry after mask redraw
Jean-Philippe Andre [Tue, 23 Dec 2014 09:57:45 +0000 (18:57 +0900)]
Evas masking: Fix invalid geometry after mask redraw

The proper geometry should be set when rendering the mask,
otherwise we can't create a new surface if it changed dimensions.

9 years agoEvas masking: Fix another set of rendering artifacts
Jean-Philippe Andre [Mon, 22 Dec 2014 12:46:14 +0000 (21:46 +0900)]
Evas masking: Fix another set of rendering artifacts

9 years agoEvas masking: Fix mask blend functions (SW engine)
Jean-Philippe Andre [Mon, 22 Dec 2014 11:29:27 +0000 (20:29 +0900)]
Evas masking: Fix mask blend functions (SW engine)

The selected op func was not performing the correct operation,
thus producing rendering artifacts. These functions should not
be used anywhere except in case of masking... which was not an
available option earlier.

It was doing (wrong):
dst = interp(mask, src, dst)

Instead of (correct):
dst = dst + (1 - mask) * src

NOTE:
This commit also disables MMX, SSE3 & NEON implementations of
pixel_mask blend operations, since they are also broken.

9 years agoEvas masking: Add rectangle masking for GL
Jean-Philippe Andre [Tue, 2 Dec 2014 06:19:10 +0000 (15:19 +0900)]
Evas masking: Add rectangle masking for GL

9 years agoEvas masking: Simplify and fix vertex logic in GL
Jean-Philippe Andre [Mon, 1 Dec 2014 13:33:59 +0000 (22:33 +0900)]
Evas masking: Simplify and fix vertex logic in GL

There was some geometry problem, and this commit also
simplifies a lot of code by factorizing it a lot.

9 years agoEvas masking: Use new texm sampler instead of texa
Jean-Philippe Andre [Mon, 1 Dec 2014 12:11:56 +0000 (21:11 +0900)]
Evas masking: Use new texm sampler instead of texa

texa should be used only for RGB+A whereas texm should
be used for masking. This should fix RGB+A masking, too.

9 years agoEvas masking: Add support for YUV, YUY2, NV12 masking in GL
Jean-Philippe Andre [Mon, 1 Dec 2014 10:52:13 +0000 (19:52 +0900)]
Evas masking: Add support for YUV, YUY2, NV12 masking in GL

RGB+A masking needs to be implemented, because there's a
clash (only one texture name "texa" is available).

9 years agoEvas masking: Add support for BGRA/ARGB masking
Jean-Philippe Andre [Mon, 1 Dec 2014 09:26:47 +0000 (18:26 +0900)]
Evas masking: Add support for BGRA/ARGB masking

Also, refactor font & image GL masking.

9 years agoEvas masking: Font masking for GL
Jean-Philippe Andre [Tue, 25 Nov 2014 05:34:42 +0000 (14:34 +0900)]
Evas masking: Font masking for GL

9 years agoEvas masking: Implement image masking for GL engines
Jean-Philippe Andre [Mon, 24 Nov 2014 11:24:25 +0000 (20:24 +0900)]
Evas masking: Implement image masking for GL engines

9 years agoEvas masking: Implement support for map draw (SW)
Jaeun Choi [Wed, 19 Nov 2014 11:08:03 +0000 (20:08 +0900)]
Evas masking: Implement support for map draw (SW)

9 years agoEvas masking: Implement support for polygon draw (SW)
Jaeun Choi [Thu, 20 Nov 2014 05:56:36 +0000 (14:56 +0900)]
Evas masking: Implement support for polygon draw (SW)

9 years agoEvas masking: Implement support for line draw (SW)
Jaeun Choi [Thu, 13 Nov 2014 11:45:14 +0000 (20:45 +0900)]
Evas masking: Implement support for line draw (SW)

Signed-off-by: Jean-Philippe Andre <jp.andre@samsung.com>
9 years agoEvas masking: Add edje support
Jaeun Choi [Mon, 17 Nov 2014 04:16:56 +0000 (13:16 +0900)]
Evas masking: Add edje support

Signed-off-by: Jean-Philippe Andre <jp.andre@samsung.com>
9 years agoEvas masking: Implement event masking support
Jaeun Choi [Thu, 13 Nov 2014 01:50:00 +0000 (10:50 +0900)]
Evas masking: Implement event masking support

In case the clipper is a mask object, we should use precise
event masking. By default precise_is_inside is not enabled
because it is expensive, but it should probably be set by
the application when they use masks as clippers.

Signed-off-by: Jean-Philippe Andre <jp.andre@samsung.com>
9 years agoEvas masking: Implement mask support in evas_render
Jean-Philippe Andre [Thu, 13 Nov 2014 01:47:29 +0000 (10:47 +0900)]
Evas masking: Implement mask support in evas_render

This implements supports for masking inside evas_render, which
means:
- Render the mask itself into a surface (ALPHA if possible)
- Pass this mask surface to the draw context
- Apply mask recursively in case a masked object is contained
  by another masked object.

@feature

9 years agoEvas masking: Allow setting an image object as clipper
Jean-Philippe Andre [Thu, 13 Nov 2014 01:40:16 +0000 (10:40 +0900)]
Evas masking: Allow setting an image object as clipper

The render function is not yet implemented, but this
prepares the infra used to set an object other than
Rectangle as a clipper.

@feature

9 years agoEvas masking: Add clip_image_[un]set functions to draw context
Jean-Philippe Andre [Thu, 13 Nov 2014 01:31:41 +0000 (10:31 +0900)]
Evas masking: Add clip_image_[un]set functions to draw context

This allows passing a mask image to the render functions.

@feature

9 years agoEvas masking: Use alpha mask in SW engine draw functions
Jaeun Choi [Thu, 13 Nov 2014 01:04:00 +0000 (10:04 +0900)]
Evas masking: Use alpha mask in SW engine draw functions

Work done by Jaeun Choi, rebased & squashed by jpeg.

This commit introduces changes to the low-level draw functions
of the SW engine considering the existence of an alpha mask image.

Features:
- Font masking (TEXT, TEXTBLOCK),
- Rectangle masking,
- Image masking (all image scaling functions should be handled).

The mask image itself is not yet set in the draw context (see
following commits).

@feature

Signed-off-by: Jean-Philippe Andre <jp.andre@samsung.com>
9 years agotests: Remove file after Evas 3D save test
Jean-Philippe Andre [Wed, 7 Jan 2015 06:01:00 +0000 (15:01 +0900)]
tests: Remove file after Evas 3D save test

This should remove src/.eet and maybe src/.ply

make distcheck would fail on my machine because src/.eet
was not removed.

9 years ago[Evas: Evas_3D] Fixed bug with bounding shapes update.
Dmytro Dadyka [Wed, 7 Jan 2015 03:58:51 +0000 (12:58 +0900)]
[Evas: Evas_3D] Fixed bug with bounding shapes update.

Reviewers: cedric, Hermet

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1851

9 years agoevas/examples: do build evas_3d_shadows
ChunEon Park [Wed, 7 Jan 2015 03:04:21 +0000 (12:04 +0900)]
evas/examples: do build evas_3d_shadows

9 years ago[Evas: Evas_3D]Fixed bug with shadows and blending of vertex position
Dmytro Dadyka [Wed, 7 Jan 2015 03:00:47 +0000 (12:00 +0900)]
[Evas: Evas_3D]Fixed bug with shadows and blending of vertex position

Reviewers: cedric, Hermet

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1850

9 years agofix build when cxx is disabled
Mike Blumenkrantz [Tue, 6 Jan 2015 20:09:11 +0000 (15:09 -0500)]
fix build when cxx is disabled

9 years agoedje shadow warning--
Mike Blumenkrantz [Tue, 6 Jan 2015 19:55:54 +0000 (14:55 -0500)]
edje shadow warning--

added in c501f2ce

ref D1768

9 years agoecore-drm: Fix issue with TakeControl and ReleaseControl functions
Chris Michael [Tue, 6 Jan 2015 18:01:31 +0000 (13:01 -0500)]
ecore-drm: Fix issue with TakeControl and ReleaseControl functions

Summary: The dbus calls to TakeControl and ReleaseControl of a session
are actual Methods that need to be setup and called in order to
operate properly. As such, this commit fixes that issue by using the
proper eldbus method calls, and fixes an issue where shutting down
Enlightenment would lead to "cannot release control" error messages.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Fix incorrect launcher shutdown procedure
Chris Michael [Tue, 6 Jan 2015 16:12:54 +0000 (11:12 -0500)]
ecore-drm: Fix incorrect launcher shutdown procedure

Summary: We need to clean tty handlers and close the tty Before we can
call logind_disconnect as that function ends up shuttting down our
dbus connection.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-evas-drm: Perform shutdown in proper order
Chris Michael [Tue, 6 Jan 2015 16:10:44 +0000 (11:10 -0500)]
ecore-evas-drm: Perform shutdown in proper order

Summary: This fixes an incorrect order with shutdown of drm library.
Sprites are created (during init) before inputs, so they should be
shutdown After inputs are.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeet: Remove unused variable
Chris Michael [Tue, 6 Jan 2015 14:05:37 +0000 (09:05 -0500)]
eet: Remove unused variable

Summary: This just removes an unused variable spotted during compile.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore_evas: fix doc for ecore_evas_pointer_warp.
Andrii Kroitor [Tue, 6 Jan 2015 11:48:33 +0000 (12:48 +0100)]
ecore_evas: fix doc for ecore_evas_pointer_warp.

Reviewers: cedric

Subscribers: cedric, reutskiy.v.v

Differential Revision: https://phab.enlightenment.org/D1815

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoedje: Edje_Edit - add Proxy to Edje_Part_Collection_Directory_Entry initializatoin.
Cedric BAIL [Tue, 6 Jan 2015 11:41:25 +0000 (12:41 +0100)]
edje: Edje_Edit - add Proxy to Edje_Part_Collection_Directory_Entry initializatoin.

Summary:
Add proxy while init Edje_Part_Collection_Directory_Entry at edje_edit_group_add and edje_edit_group_alias_add.

@fix

@fix

Reviewers: Hermet, seoz, cedric, raster

Subscribers: reutskiy.v.v

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D1846

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeet: refactoring jpeg encoding.
Cedric BAIL [Tue, 6 Jan 2015 10:42:55 +0000 (11:42 +0100)]
eet: refactoring jpeg encoding.

9 years agoeet: refactorize handling of endian.
Cedric BAIL [Tue, 6 Jan 2015 10:09:27 +0000 (11:09 +0100)]
eet: refactorize handling of endian.

9 years agoeet: use eina_swap*() function instead of custom slower one.
Cedric BAIL [Fri, 3 Jan 2014 07:20:53 +0000 (16:20 +0900)]
eet: use eina_swap*() function instead of custom slower one.

9 years agoEvas fonts: Fix minor deviation in RLE font render
Jean-Philippe Andre [Tue, 6 Jan 2015 10:15:07 +0000 (19:15 +0900)]
Evas fonts: Fix minor deviation in RLE font render

So I've discovered some weird output values after drawing
some text. The destination alpha would become 0xFE even
when the back buffer had a background with 0xFF alpha.

Example:
Dest is 0xff00ff00 (green).
Color is 0xffffffff (white).
Current font alpha is 170 (0xaa).
--> Output was 0xFEaaFEaa instead of 0xFFaaFFaa.

This is because of some slightly invalid calculation
when doing the font masking (mtab[v] = 0x55 above).

Indeed, MUL_256 takes alpha values in the range [1-256]
and not [0-256] as was assumed.

9 years agoeolian-cxx: Fixed mixed inheritance between C and C++ classes
Felipe Magno de Almeida [Mon, 5 Jan 2015 20:59:09 +0000 (18:59 -0200)]
eolian-cxx: Fixed mixed inheritance between C and C++ classes

Removed genereation for Eo.Base primitive methods which segfault'ed
on parent_set function.

9 years agoeolian_cxx: Using eina::optional to handle parameters without @nonull property
Vitor Sousa [Mon, 5 Jan 2015 17:41:37 +0000 (15:41 -0200)]
eolian_cxx: Using eina::optional to handle parameters without @nonull property

9 years agoeina_cxx: Allow creating an eina::optional<T> from a type convertible to T
Vitor Sousa [Mon, 5 Jan 2015 17:35:49 +0000 (15:35 -0200)]
eina_cxx: Allow creating an eina::optional<T> from a type convertible to T

9 years agoeo_cxx: Implicit conversion from C++ wrapper to void* via address_of operator
Vitor Sousa [Mon, 5 Jan 2015 17:26:11 +0000 (15:26 -0200)]
eo_cxx: Implicit conversion from C++ wrapper to void* via address_of operator

9 years agoeolian_cxx: Fix compilation of inheritance from Eolian generated classes
Vitor Sousa [Tue, 30 Dec 2014 23:18:40 +0000 (21:18 -0200)]
eolian_cxx: Fix compilation of inheritance from Eolian generated classes

Updated the code for the "inheritance helper" functions and classes
to be in conformance with the new wrapper architecture.

Member variable "parents" of efl::eo::eo_class struct was split in two
member variables, "ancestors" and "parents", the former containing all
Eo class ancestors and the later only the direct bases.
Changed all required files accordingly.

Check to avoid using biding string when the parameter is @out now matches
any variation of "char".

Add default constructor to efl::eo::concrete in order to allow it to be
NULL initialized in the "inheritance helper" classes.

Removed conflicting parent_set member function in the efl::eo::inherit
class.

Removed the "inheritance_extension_function" generator since it is no
longer used.

9 years agoallow C++ user create new classes inheriting from Eolian generated Classes
Larry Jr [Tue, 26 Aug 2014 18:54:28 +0000 (15:54 -0300)]
allow C++ user create new classes inheriting from Eolian generated Classes

9 years agoeolian_cxx: Add implicit conversion of the returned pointer of C++ wrappers
Vitor Sousa [Tue, 23 Dec 2014 19:40:11 +0000 (17:40 -0200)]
eolian_cxx: Add implicit conversion of the returned pointer of C++ wrappers

Overloaded address-of operator for C++ Eolian wrappers for implicit converting
to any ancestor pointer type when getting the object memory address.

Added new grammars to aid the creation of the pointer wrappers
responsible of doing the implicit conversions.

Added a unit test which checks the address-of overload for arbitrary classes.
Added new .eo files to be used in it.

9 years agoeolian_cxx: Fix "dreaded diamond" inheritance problem for C++ wrappers
Vitor Sousa [Fri, 19 Dec 2014 16:34:15 +0000 (14:34 -0200)]
eolian_cxx: Fix "dreaded diamond" inheritance problem for C++ wrappers

Solved diamond inheritance problem by completely removing inheritance in
the abstract class.
All ancestors are inherited directly in the concrete class.
The algorithm that list the ancestors also avoid repetition.

Now concrete classes define methods too. This helps referring the correct
method directly by the object type (when there are methods with the same
name).

Moved the declaration and definition of constructor methods to the concrete
class, since they should not be used in derived classes.

Updated example that call "color_set".
With this model, if two ancestor classes have a method with the same name,
to call one of them from a derived class you must write the scoped name of
the member function in the wrapper.
In this case, either Evas.Object and Evas.SmartObject have a property
named "color".

Added "from_global" option to the full_name grammar too.

9 years agocxx: General C++ Eo wrapper generation improvement
Vitor Sousa [Wed, 17 Dec 2014 13:47:19 +0000 (11:47 -0200)]
cxx: General C++ Eo wrapper generation improvement

Using a new architecture in the generated files that simplify multiple
inheritance and allows the use of interface types as parameters.

No longer using a hand-crafted C++ header for eo_base.eo.
This file was added to the generation process.

Updated all files that are dependent in the hand-crafted eo_base
C++ header.

Now there is a class that contains the essentials functions of the former
eo::base wrapper and that is used to create (through inheritance) the
"concrete" classes for all Eo generated wrappers.

No longer binding any function or property that are protected, private or
legacy for now.

eolian_type_instance is now a struct with general information for the
whole type.

Added the new header file namespace_generator.hh to hold namespace
generation grammars.

Separated declaration and definition of Eo wrappers methods.

Referring for most objects by its full name (starting at the global
namespace ::) in the generated files.

Created additional helper grammars to avoid code replication.

Removed a TODO comment referring to a doubt about inheritance of
constructor methods.
Added a TODO comment regarding memory allocation for callbacks in static
member functions.

9 years agocatch exceptions in cxx wrappers and throw a eina_error
Larry Jr [Thu, 14 Aug 2014 18:20:49 +0000 (15:20 -0300)]
catch exceptions in cxx wrappers and throw a eina_error

9 years agoeolian_cxx: Fix: Using binding type for @out parameter instead of native type
Vitor Sousa [Tue, 25 Nov 2014 18:18:08 +0000 (16:18 -0200)]
eolian_cxx: Fix: Using binding type for @out parameter instead of native type

No longer reverting to the native type when the parameter has "@out"
direction.

Added "is_out" member variable to eolian_type class.
With that, generators can keep track of the direction of the parameters.
Also added helper functions "type_is_out" and "type_is_complex".

Created "to_native" functions in eo_cxx_interop.hh to convert binding types
from C++ arguments to the actual C function arguments.
Added static assertions in these functions to enforce compatibility
between the binding and the native type (Required by @out parameters).

Reworked the overload of the "to_c" function for eo::base derivated
objects. Now there is a overload that rely in the compatibility
between the native type and the wrapper, enabling a wrapper to be used as
an output parameter.

9 years agoeolian-cxx: Fixed C++ conversion for lists
Felipe Magno de Almeida [Sun, 26 Oct 2014 21:42:55 +0000 (19:42 -0200)]
eolian-cxx: Fixed C++ conversion for lists

Signed-off-by: Vitor Sousa <vitorsousasilva@gmail.com>
9 years agoadded new types in lookup table
Larry Jr [Tue, 7 Oct 2014 18:06:01 +0000 (15:06 -0300)]
added new types in lookup table

Signed-off-by: Vitor Sousa <vitorsousasilva@gmail.com>