platform/upstream/efl.git
9 years agoeina - begin event log infra we can get from the new debug monitor
Carsten Haitzler (Rasterman) [Sun, 10 May 2015 10:05:54 +0000 (19:05 +0900)]
eina - begin event log infra we can get from the new debug monitor

we can down dump event logs. some ecore mainloop bits are logging at
the moment.

9 years agoevas/engine: Fix the ector drawing for gl backend
Subhransu Mohanty [Sat, 9 May 2015 09:46:42 +0000 (18:46 +0900)]
evas/engine: Fix the ector drawing for gl backend

9 years agoevas/canvas : Call correct engine function for setting clip color
Subhransu Mohanty [Sat, 9 May 2015 09:44:31 +0000 (18:44 +0900)]
evas/canvas : Call correct engine function for setting clip color

9 years agoEcore tests: REMOVE wayland tests.
Tom Hacohen [Fri, 8 May 2015 15:17:05 +0000 (16:17 +0100)]
Ecore tests: REMOVE wayland tests.

These tests aren't really testing anything, and are just broken and annoying.
They only test init and shutdown, and they require a running wayland compositor.

They fail for anyone that has wayland enabled but not running in a wayland session.

They should be brought back once they actual test something, or once we allow skipping
tests that can't be run due to environment issues.

9 years agoEcore tests: Properly split (some) unit tests to units
Tom Hacohen [Fri, 8 May 2015 14:23:25 +0000 (15:23 +0100)]
Ecore tests: Properly split (some) unit tests to units

Units were not properly split. We also had to resort to hacks we are not
allowed to do before, which are now fixed.
We don't allow shutting down and powering up again. Once something is off,
it's off.

9 years agoEo: Mark composite APIs as beta.
Tom Hacohen [Fri, 8 May 2015 13:21:23 +0000 (14:21 +0100)]
Eo: Mark composite APIs as beta.

Until now we used @protected, but now we can finally properly use @beta.

9 years agoEcore audio test: Remove unneeded test.
Tom Hacohen [Fri, 8 May 2015 13:04:00 +0000 (14:04 +0100)]
Ecore audio test: Remove unneeded test.

We don't need to test eo works in the ecore audio suite.
If you mark it as virtual, it's virtual.

9 years agoeolian: generation for beta events
Daniel Kolesa [Fri, 8 May 2015 13:45:10 +0000 (14:45 +0100)]
eolian: generation for beta events

9 years agoeolian: API and tests for beta events
Daniel Kolesa [Fri, 8 May 2015 13:37:02 +0000 (14:37 +0100)]
eolian: API and tests for beta events

9 years agoeolian: fix generation tests
Daniel Kolesa [Fri, 8 May 2015 13:28:42 +0000 (14:28 +0100)]
eolian: fix generation tests

9 years agoeolian: oops... forgot to add a param
Daniel Kolesa [Fri, 8 May 2015 13:23:23 +0000 (14:23 +0100)]
eolian: oops... forgot to add a param

9 years agoeolian: parsing for beta events
Daniel Kolesa [Fri, 8 May 2015 13:22:32 +0000 (14:22 +0100)]
eolian: parsing for beta events

9 years agoeolian: beta funcs generation
Daniel Kolesa [Fri, 8 May 2015 13:17:20 +0000 (14:17 +0100)]
eolian: beta funcs generation

9 years agobuild: Pass on makeflags and use proper @$(MAKE)
Stefan Schmidt [Fri, 8 May 2015 13:08:44 +0000 (15:08 +0200)]
build: Pass on makeflags and use proper @$(MAKE)

Going with the build in automake stuff allows us to run this target also in
parallel. Before we got the warning that the jobserver was missing and it
would only run as -j1. Thanks to Tom for spotting and Flameeyes for the blog
post:
https://blog.flameeyes.eu/2010/10/tell-tale-signs-that-your-makefile-is-broken#gsc.tab=0

9 years agoeolian: is_beta API plus tests
Daniel Kolesa [Fri, 8 May 2015 13:12:27 +0000 (14:12 +0100)]
eolian: is_beta API plus tests

9 years agoeolian: parse the @beta qualifier
Daniel Kolesa [Fri, 8 May 2015 13:00:07 +0000 (14:00 +0100)]
eolian: parse the @beta qualifier

9 years agoEfl debug: Rename shadowing variable.
Tom Hacohen [Fri, 8 May 2015 12:55:39 +0000 (13:55 +0100)]
Efl debug: Rename shadowing variable.

9 years agolibunibreak: Follow README rename in EXTRA_DIST for distcheck.
Stefan Schmidt [Fri, 8 May 2015 12:29:13 +0000 (14:29 +0200)]
libunibreak: Follow README rename in EXTRA_DIST for distcheck.

The latest libunibreak updated changed the README name to README.md. Follow
this for the EXTRA_DIST files.

9 years agoautotools: enable make check-build which builds only test suites
kabeer khan [Fri, 8 May 2015 12:19:48 +0000 (14:19 +0200)]
autotools: enable make check-build which builds only test suites

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

Reviewers: tasn, cedric, stefan_schmidt

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2480

9 years agoevas vector: remove white trailings
ChunEon Park [Fri, 8 May 2015 11:51:52 +0000 (20:51 +0900)]
evas vector: remove white trailings

9 years agoevas/canvas: Null check for base class data pointer.
Subhransu Mohanty [Fri, 8 May 2015 11:32:40 +0000 (20:32 +0900)]
evas/canvas: Null check for base class data pointer.

9 years agoevas vector: cancel the transformation when null is passed.
ChunEon Park [Thu, 7 May 2015 11:02:01 +0000 (20:02 +0900)]
evas vector: cancel the transformation when null is passed.

9 years agoefl debug - clean up debugd and debug cli tool code proto handling
Carsten Haitzler (Rasterman) [Fri, 8 May 2015 10:48:26 +0000 (19:48 +0900)]
efl debug - clean up debugd and debug cli tool code proto handling

this clenas up protocol handling to share common code and have more
compact and easier to maintain code on both sides here.

9 years agoeina debug - add more comments to debugd connection
Carsten Haitzler (Rasterman) [Fri, 8 May 2015 08:23:38 +0000 (17:23 +0900)]
eina debug - add more comments to debugd connection

9 years agoeina debug - add comments
Carsten Haitzler (Rasterman) [Fri, 8 May 2015 08:18:38 +0000 (17:18 +0900)]
eina debug - add comments

9 years agoefl debug - start on common file for efl_debug tools for protocol
Carsten Haitzler (Rasterman) [Fri, 8 May 2015 08:14:24 +0000 (17:14 +0900)]
efl debug - start on common file for efl_debug tools for protocol

this should reduce duplication between debugd and debug tools

9 years agoeina_ddebug: Fix build for the case without all requirements for eina_debug
Stefan Schmidt [Fri, 8 May 2015 07:55:12 +0000 (09:55 +0200)]
eina_ddebug: Fix build for the case without all requirements for eina_debug

In eina_debug.h we only include the eina headsers when all defines are true
for backtrace, dlopen, unwind, etc. We still use Eina_Bool here when these
cases are not met. I just hit this case and failed the build.

9 years agoeina debug - clean up protocol handling inside eina
Carsten Haitzler (Rasterman) [Fri, 8 May 2015 07:58:18 +0000 (16:58 +0900)]
eina debug - clean up protocol handling inside eina

9 years agoeina - start a much improved eina dbug infra and have eina_log use it
Carsten Haitzler (Rasterman) [Tue, 5 May 2015 02:35:16 +0000 (11:35 +0900)]
eina - start a much improved eina dbug infra and have eina_log use it

this makes eina_log give bt's for all error logs. this is very useful
in finding just where a problem happens. the problem int he past is
that these have not been too useful due to backtrace_symbols() being
"useless". thus use the eina_btlog tool i added too.

also started infra for a debug monitor that can use the backtrace
infra to collect runtime stats ANY TIME for a process (don't need to
run under a debugger).

@feat

9 years agoecore - ecore_pipe_wait timeout broken as it never adjusts while waiting
Carsten Haitzler (Rasterman) [Tue, 5 May 2015 10:28:14 +0000 (19:28 +0900)]
ecore - ecore_pipe_wait timeout broken as it never adjusts while waiting

if select gets interrupted it just waits again from the start because
it uses loop time not "now" time. this is wrong and makes timeout
waits possibly hang if enough things interrupt select without reading
data. this fixes that.

@fix

9 years agoefl - emotion modules - unbreak due to EAPI changes
Carsten Haitzler (Rasterman) [Fri, 8 May 2015 05:12:29 +0000 (14:12 +0900)]
efl - emotion modules - unbreak due to EAPI changes

emotion relied on Emotion.h's EAPI defs... now they have been removed
fix!

9 years agoMerge branch 'devs/devilhorns/ecore_drm'
Chris Michael [Thu, 7 May 2015 19:40:38 +0000 (15:40 -0400)]
Merge branch 'devs/devilhorns/ecore_drm'

Summary: Merge new 1.15 API functions for ecore_drm which will be used
in the E RandR config dialog.

NB: This is basically a set of API functions that are needed for
working with the new e_randr2 codebase in E. This makes the dialog
'usable' in E, however there is no API yet for "applying" those
settings changes. That will come shortly.

@feature

9 years agoRevert "autotools: enable make check per individual modules."
Stefan Schmidt [Thu, 7 May 2015 18:50:56 +0000 (20:50 +0200)]
Revert "autotools: enable make check per individual modules."

This reverts commit 35119e7bfdc7c13c2041293f3d0b2ebe1fb7c313.

Reverted to bring make check back in a working state. Also the way we
want to handle a more modular testing needs discussion.

9 years agoecore-drm: Add new API function to mark an output as primary
Chris Michael [Thu, 7 May 2015 15:37:54 +0000 (11:37 -0400)]
ecore-drm: Add new API function to mark an output as primary

Summary: This adds a new API function which we can use to mark a given
Ecore_Drm_Output as being the primary output.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to find an output given a name
Chris Michael [Thu, 7 May 2015 15:31:46 +0000 (11:31 -0400)]
ecore-drm: Add API function to find an output given a name

Summary: This adds a new API function to find an Ecore_Drm_Output
which matches a given name.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Fix formatting
Chris Michael [Thu, 7 May 2015 13:24:15 +0000 (09:24 -0400)]
ecore-drm: Fix formatting

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to get an output's crtc size
Chris Michael [Wed, 6 May 2015 17:19:08 +0000 (13:19 -0400)]
ecore-drm: Add API function to get an output's crtc size

Summary: This adds a new API function to return an output's crtc size.
This is mainly used for drm RandR config in E

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return primary output
Chris Michael [Wed, 6 May 2015 16:22:15 +0000 (12:22 -0400)]
ecore-drm: Add API function to return primary output

Summary: This adds a new API function to return the output which is
marked as the primary output.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add support for a Primary Output
Chris Michael [Wed, 6 May 2015 16:15:56 +0000 (12:15 -0400)]
ecore-drm: Add support for a Primary Output

Summary: As 'primary' output support is not implemented in hardware,
we need to support this feature via software. For now, the first
output returned via libdrm will be marked as 'primary' until user
changes it via config

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Set drm client capabilities to support universal planes
Chris Michael [Wed, 6 May 2015 15:31:57 +0000 (11:31 -0400)]
ecore-drm: Set drm client capabilities to support universal planes

Summary: If we do not set client capabilities to support universal
planes, then libdrm does not expose the primary or cursor planes so
this adds a quick call to set DRM_CLIENT_CAP_UNIVERSAL_PLANES

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Cleanup plane debug output
Chris Michael [Wed, 6 May 2015 15:31:25 +0000 (11:31 -0400)]
ecore-drm: Cleanup plane debug output

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Set output connected property in the creation function
Chris Michael [Wed, 6 May 2015 13:41:03 +0000 (09:41 -0400)]
ecore-drm: Set output connected property in the creation function

Summary: Makes more sense to set the output's connected property
inside the function which creates outputs.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add some debug code to spit out planes and plane properties
Chris Michael [Wed, 6 May 2015 13:33:17 +0000 (09:33 -0400)]
ecore-drm: Add some debug code to spit out planes and plane properties

Summary: This adds a debug function to spit out planes and their
properties while debugging rotation support

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return output modes
Chris Michael [Tue, 5 May 2015 13:12:25 +0000 (09:12 -0400)]
ecore-drm: Add API function to return output modes

Summary: This adds a new API function to return a list of modes
supported on a given output

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return the edid of a given output
Chris Michael [Mon, 4 May 2015 19:03:47 +0000 (15:03 -0400)]
ecore-drm: Add API function to return the edid of a given output

Summary: This adds a new API function to return the EDID string of a
given output.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return if a given output has a backlight
Chris Michael [Mon, 4 May 2015 18:41:18 +0000 (14:41 -0400)]
ecore-drm: Add API function to return if a given output has a backlight

Summary: This adds a new API function to call so we can check if a
given Ecore_Drm_Output has a backlight

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return the connector type of a given output
Chris Michael [Mon, 4 May 2015 18:26:47 +0000 (14:26 -0400)]
ecore-drm: Add API function to return the connector type of a given output

Summary: This adds a new API function to return the connector type of
an Ecore_Drm_Output

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add connector type to output structure
Chris Michael [Mon, 4 May 2015 18:22:10 +0000 (14:22 -0400)]
ecore-drm: Add connector type to output structure

Summary: This adds a connector type field to the Output structure.
This will be used in the Enlightenment Drm rendering code for RandR
config

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return output connected state
Chris Michael [Mon, 4 May 2015 18:13:21 +0000 (14:13 -0400)]
ecore-drm: Add API function to return output connected state

Summary: This adds a new API function to check if a given output is connected or not.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return the name of an output
Chris Michael [Mon, 4 May 2015 18:06:31 +0000 (14:06 -0400)]
ecore-drm: Add API function to return the name of an output

Summary: This adds a new API function to turn the name of a given Ecore_Drm_Output.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Add API function to return screen size range
Chris Michael [Mon, 4 May 2015 17:25:15 +0000 (13:25 -0400)]
ecore-drm: Add API function to return screen size range

Summary: This adds new API function to return the min & max width &
height of valid screen range for an Ecore_Drm_Device

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeolian: remove old property syntax support
Daniel Kolesa [Thu, 7 May 2015 16:35:33 +0000 (17:35 +0100)]
eolian: remove old property syntax support

9 years agoeolian: change all EFL .eo files to use new syntax for properties
Daniel Kolesa [Thu, 7 May 2015 15:36:45 +0000 (16:36 +0100)]
eolian: change all EFL .eo files to use new syntax for properties

9 years agoeolian: allow the new property syntax
Daniel Kolesa [Thu, 7 May 2015 13:57:46 +0000 (14:57 +0100)]
eolian: allow the new property syntax

9 years agoCheck a key repeat using xkb ruled keycode
jhyuni.kang [Thu, 7 May 2015 13:24:54 +0000 (09:24 -0400)]
Check a key repeat using xkb ruled keycode

Summary:
ecore_wayland get a key information using libxkbcommon's functions.
         Also generate key repeat check xkb keymap's information using xkb_keymap_key_repeats().
         But xkb rules reflect X broken keycodes, so offset by 8.
         In ecore_wayland backend, other xkb funtions using offset keycode
         but only use default keycode to check key repeat.
         So I fix it to use xkb ruled keycode.

Test Plan: Long press a key and watch repeated key events generate well

Reviewers: devilhorns, raster

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2475

9 years agoelua: use a minimal app to avoid loading eolian at runtime
Daniel Kolesa [Thu, 7 May 2015 10:56:59 +0000 (11:56 +0100)]
elua: use a minimal app to avoid loading eolian at runtime

Otherwise it would fail on systems with no Eolian installed.

@fix

9 years agoeolian: reserve @beta and @property
Daniel Kolesa [Thu, 7 May 2015 10:35:14 +0000 (11:35 +0100)]
eolian: reserve @beta and @property

9 years agoStatic deps unibreak: update to what will soon be version 3.
Tom Hacohen [Thu, 7 May 2015 09:53:11 +0000 (10:53 +0100)]
Static deps unibreak: update to what will soon be version 3.

Version 3 is not yet released, but this is on track to become it.
This is based on commit: a815e11f7ebf35b59278f783227a829ee4692760.

@feature.

9 years agoRevert "Static deps: Move unibreak to be an external dep."
Tom Hacohen [Thu, 7 May 2015 09:40:57 +0000 (10:40 +0100)]
Revert "Static deps: Move unibreak to be an external dep."

Apparently the Debian package, while up to date, for some reason does
not ship the .pc file (according to q66).
According to Stefan, Fedora doesn't even have libunibreak, but only the
previous naming and old version.

Will have to wait a few years more. :(

This reverts commit a2a9f33802a3923c1469789f66d5fdab1eaea943.

9 years agolz4: Update our internal copy to release r128
Stefan Schmidt [Thu, 7 May 2015 08:52:54 +0000 (10:52 +0200)]
lz4: Update our internal copy to release r128

Looking through the git log it is unclear which release we used before as nobody
stated it there. :/ We updated after the security issues last year so my best
guess is that we have something like r119.

To see what changed I now included the NEWS file and also the LICENSE file from
upstream. Upstream in now hosted here: https://github.com/Cyan4973/lz4 and
http://www.lz4.info

I recommend STRONGLY that you check if your distro ships liblz4 as an up to
date library package and use the --enable-liblz4 configure option to use the
system version. I consider making the system version default for upcoming
releases and only carry the internal one as fallback for systems that do not
provide it.

Fix T2374

9 years agoeet: Remove direct usage of lz4 as we use emile for this now.
Stefan Schmidt [Thu, 7 May 2015 08:28:58 +0000 (10:28 +0200)]
eet: Remove direct usage of lz4 as we use emile for this now.

Left over from the point where it got migrated to emile.

9 years agobuild: Do not show mkdir calls in silent mode
Stefan Schmidt [Thu, 7 May 2015 07:51:58 +0000 (09:51 +0200)]
build: Do not show mkdir calls in silent mode

Keep this noise out of the build ouput. Still visible in verbose mode.

9 years agodocs: Fix access to main elocation page from front page
Stefan Schmidt [Thu, 7 May 2015 07:38:12 +0000 (09:38 +0200)]
docs: Fix access to main elocation page from front page

9 years agoStatic deps: Move unibreak to be an external dep.
Tom Hacohen [Thu, 7 May 2015 09:02:40 +0000 (10:02 +0100)]
Static deps: Move unibreak to be an external dep.

We need any version of libunibreak. The first one has been released in mid 2012.
Even slow distros like ubuntu already have an LTS out with a good enough
version, so I consider this enough to remove the maintenance cost.
This has been discussed on IRC.

@feature

9 years agoevil: fix dladdr() implementation, add 2 unit tests for evil.
Vincent Torri [Wed, 29 Apr 2015 08:50:11 +0000 (10:50 +0200)]
evil: fix dladdr() implementation, add 2 unit tests for evil.

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevil: fix mem leak in dlsym() when using UNICODE
Vincent Torri [Sun, 26 Apr 2015 14:38:10 +0000 (16:38 +0200)]
evil: fix mem leak in dlsym() when using UNICODE

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevil: minor formatting in dlfcn.c
Vincent Torri [Sun, 26 Apr 2015 14:36:08 +0000 (16:36 +0200)]
evil: minor formatting in dlfcn.c

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: do not cast uint64x1_t to int in evas_common_convert_argb_premul
Cedric BAIL [Tue, 28 Apr 2015 21:40:58 +0000 (23:40 +0200)]
evas: do not cast uint64x1_t to int in evas_common_convert_argb_premul

Summary:
In GCC 5.1 arm_neon header for aarch64 was changed. It is not possible anymore to silently cast uint64x1_t to int.

So replace cast with proper getter function to avoid following error:
lib/evas/common/evas_convert_color.c:50:18: error: incompatible types when assigning to type 'DATA32 {aka unsigned int}' from type 'uint64x1_t'
nas += vpaddl_u32(vpaddl_u16(vpaddl_u8(cmp)));

Reviewers: raster, cedric, devilhorns

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2443

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: implement _op_blend_rel_mas_c_dp_neon using NEON intrinsics
Cedric BAIL [Tue, 28 Apr 2015 21:40:04 +0000 (23:40 +0200)]
evas: implement _op_blend_rel_mas_c_dp_neon using NEON intrinsics

Summary: NEON intrinsics can be built both for armv7 and armv8.

Reviewers: raster, cedric

Reviewed By: cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2442

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: implement _op_blend_rel_{p,pan}_dp_neon using NEON intrinsics
Cedric BAIL [Tue, 28 Apr 2015 21:39:18 +0000 (23:39 +0200)]
evas: implement _op_blend_rel_{p,pan}_dp_neon using NEON intrinsics

Summary: NEON intrinsics can be built both for armv7 and armv8.

Reviewers: raster, cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2441

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: implement _op_blend_rel_c_dp_neon using NEON intrinsics
Cedric BAIL [Tue, 28 Apr 2015 21:38:34 +0000 (23:38 +0200)]
evas: implement _op_blend_rel_c_dp_neon using NEON intrinsics

Summary: NEON intrinsics can be built both for armv7 and armv8.

Reviewers: raster, cedric

Reviewed By: cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2440

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: implement _op_blend_rel_p_c_dp_neon using NEON intrinsics
Cedric BAIL [Tue, 28 Apr 2015 21:37:37 +0000 (23:37 +0200)]
evas: implement _op_blend_rel_p_c_dp_neon using NEON intrinsics

Summary:
NEON intrinsics can be built both for armv7 and armv8.

There were no NEON variant for this function, so it was added with all copies to init function.

Reviewers: raster, cedric

Reviewed By: cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2417

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: implement pixel_color blending functions using NEON intrinsics.
Cedric BAIL [Tue, 28 Apr 2015 21:36:04 +0000 (23:36 +0200)]
evas: implement pixel_color blending functions using NEON intrinsics.

Summary:
NEON intrinsics can be built both for armv7 and armv8.
Implemented functions:
_op_blend_pan_c_dp_neon
_op_blend_p_can_dp_neon
_op_blend_pan_can_dp_neon
_op_blend_p_caa_dp_neon
_op_blend_pan_caa_dp_neon

Reviewers: raster, cedric

Subscribers: cedric

Projects: #efl

Maniphest Tasks: T2341

Differential Revision: https://phab.enlightenment.org/D2409

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add test case for eina_hash_crc function.
vivek [Tue, 28 Apr 2015 21:34:34 +0000 (23:34 +0200)]
eina: add test case for eina_hash_crc function.

Summary:
Added test case for eina_hash_crc function which uses crc32 for hashing
algorithm

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

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2296

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add crc hash function to eina hash module.
vivek [Tue, 28 Apr 2015 21:33:56 +0000 (23:33 +0200)]
eina: add crc hash function to eina hash module.

Summary:
Added eina_hash_crc function in eina hash module to generate hash code
using crc-32.

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

Subscribers: Sergeant_Whitespace, cedric

Differential Revision: https://phab.enlightenment.org/D2273

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add benchmark for crc hash in eina benchmark
vivek [Tue, 28 Apr 2015 21:29:18 +0000 (23:29 +0200)]
eina: add benchmark for crc hash in eina benchmark

Summary:
The code added is actually benchmarking all hash function key with various length (8, 32, 256 bytes).

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

Reviewed By: cedric

Subscribers: Sergeant_Whitespace, cedric

Differential Revision: https://phab.enlightenment.org/D2217

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevil: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:52:57 +0000 (16:52 +0200)]
evil: remove the need to order the header correctly for Windows.

9 years agoethumb: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:52:48 +0000 (16:52 +0200)]
ethumb: remove the need to order the header correctly for Windows.

9 years agoephysics: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:52:32 +0000 (16:52 +0200)]
ephysics: remove the need to order the header correctly for Windows.

9 years agoeolian: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:52:22 +0000 (16:52 +0200)]
eolian: remove the need to order the header correctly for Windows.

9 years agoeo: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:52:13 +0000 (16:52 +0200)]
eo: remove the need to order the header correctly for Windows.

9 years agoemotion: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:51:57 +0000 (16:51 +0200)]
emotion: remove the need to order the header correctly for Windows.

9 years agoembryo: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 14:51:44 +0000 (16:51 +0200)]
embryo: remove the need to order the header correctly for Windows.

9 years agoelua: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:28:24 +0000 (15:28 +0200)]
elua: remove the need to order the header correctly for Windows.

9 years agoeldbus: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:13:48 +0000 (15:13 +0200)]
eldbus: remove the need to order the header correctly for Windows.

9 years agoeio: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:12:42 +0000 (15:12 +0200)]
eio: remove the need to order the header correctly for Windows.

9 years agoefreet: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:12:03 +0000 (15:12 +0200)]
efreet: remove the need to order the header correctly for Windows.

9 years agoefl: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:11:20 +0000 (15:11 +0200)]
efl: remove the need to order the header correctly for Windows.

9 years agoeeze: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:09:59 +0000 (15:09 +0200)]
eeze: remove the need to order the header correctly for Windows.

9 years agoedje: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 13:09:11 +0000 (15:09 +0200)]
edje: remove the need to order the header correctly for Windows.

9 years agoecore: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 12:24:48 +0000 (14:24 +0200)]
ecore: remove the need to order the header correctly for Windows.

9 years agoevas: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 12:24:11 +0000 (14:24 +0200)]
evas: remove the need to order the header correctly for Windows.

9 years agoeet: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 12:23:37 +0000 (14:23 +0200)]
eet: remove the need to order the header correctly for Windows.

9 years agoemile: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 12:22:38 +0000 (14:22 +0200)]
emile: remove the need to order the header correctly for Windows.

9 years agoeina: remove the need to order the header correctly for Windows.
Cedric BAIL [Tue, 28 Apr 2015 12:21:27 +0000 (14:21 +0200)]
eina: remove the need to order the header correctly for Windows.

9 years agoeina: add test case for eina_crc function in eina module.
vivek [Wed, 22 Apr 2015 13:37:32 +0000 (15:37 +0200)]
eina: add test case for eina_crc function in eina module.

Summary:
I added test cases using seed 0xffffffff, I was not
able to get crc value for different seeds online. Checked some of the links, but they are
using entirely different logic and value was not matching for other seed value.

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

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2293

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoeina: add CRC implementation to Eina module.
vivek [Wed, 22 Apr 2015 13:34:33 +0000 (15:34 +0200)]
eina: add CRC implementation to Eina module.

Summary:
Added eina_crc function in eina to calculate crc for the key passed and
added eina_hash_crc function for hashing using crc

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

Reviewed By: cedric

Subscribers: Sergeant_Whitespace, cedric

Differential Revision: https://phab.enlightenment.org/D2260

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: primitives - Changed enum for types of Evas_3D_Mesh primitives
Bogdan Devichev [Wed, 22 Apr 2015 13:27:49 +0000 (15:27 +0200)]
evas: primitives - Changed enum for types of Evas_3D_Mesh primitives

Summary:
**`Creating API for primitives.`**

//What should be in the end://

- Class evas_3d_primitive inherited from eo.base
- Unificated API of mesh for setting data from primitives
- API's of primitives to get/set precision, ratio, type, tex_scale etc.
- code of primitives in separated modules

//Steps://

**1.) Enum for types of primitives**
2.) Copies of primitives in modules
3.) Eo and c files for class of primitive
4.) In common create picker of module
5.) API function in mesh.c
6.) Binding of new realization
7.) Rewriting of examples
8.) Deleting of old files for primitives

//Also this commit will be usable in parsing EDC by edje//

Reviewers: Hermet, raster, cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D2359

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: implement _op_blend_p_mas_dp_neon and _op_blend_pas_mas_dp_neon in NEON intrinsics.
Cedric BAIL [Wed, 22 Apr 2015 13:24:13 +0000 (15:24 +0200)]
evas: implement _op_blend_p_mas_dp_neon and _op_blend_pas_mas_dp_neon in NEON intrinsics.

Reviewers: raster, cedric

Reviewed By: cedric

Subscribers: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2391

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: improve _op_blend_p_dp_neon intrinsics implementation
Cedric BAIL [Wed, 22 Apr 2015 13:20:22 +0000 (15:20 +0200)]
evas: improve _op_blend_p_dp_neon intrinsics implementation

Summary:
Use vceqq and vbsl instead of twice as much vmovl and vadd instructions.
Replace vaddq_u8 with vaddq_u32.
This allows NEON code to behave exactly like C version.

Reviewers: raster, cedric

Reviewed By: cedric

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D2361

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