platform/upstream/efl.git
8 years agoefl model: add missing since tags for new APIs
Stefan Schmidt [Wed, 6 Jan 2016 14:26:30 +0000 (15:26 +0100)]
efl model: add missing since tags for new APIs

8 years agoedje: add missing since tags to new edje_size_class symbols
Stefan Schmidt [Wed, 6 Jan 2016 14:19:28 +0000 (15:19 +0100)]
edje: add missing since tags to new edje_size_class symbols

8 years agoecore_wl2: add since tag to all new symbols in 1.17 missing it
Stefan Schmidt [Wed, 6 Jan 2016 14:08:13 +0000 (15:08 +0100)]
ecore_wl2: add since tag to all new symbols in 1.17 missing it

8 years agoecore_evas: add missing since tags for new APIs coming with 1.17
Stefan Schmidt [Wed, 6 Jan 2016 13:52:56 +0000 (14:52 +0100)]
ecore_evas: add missing since tags for new APIs coming with 1.17

8 years agoEvas: Fix the build... I hope
Jean-Philippe Andre [Wed, 6 Jan 2016 08:18:57 +0000 (17:18 +0900)]
Evas: Fix the build... I hope

I forgot to modify some lines. Of course randomness made things
work for me.

8 years agoEctor: Do not install any header file
Jean-Philippe Andre [Wed, 6 Jan 2016 07:43:46 +0000 (16:43 +0900)]
Ector: Do not install any header file

Also fix spelling in .pc file

8 years agoEctor: Move ector_buffer.h along ector_surface.h
Jean-Philippe Andre [Wed, 6 Jan 2016 07:32:51 +0000 (16:32 +0900)]
Ector: Move ector_buffer.h along ector_surface.h

This makes it an installed header for now.
I still believe no Ector header should be installed.

8 years agoEvas: Do not install evas_ector_buffer
Jean-Philippe Andre [Wed, 6 Jan 2016 07:26:01 +0000 (16:26 +0900)]
Evas: Do not install evas_ector_buffer

This is an internal API, should not be installed.

8 years agoEctor GL: Fix warning with clang
Jean-Philippe Andre [Wed, 6 Jan 2016 05:31:57 +0000 (14:31 +0900)]
Ector GL: Fix warning with clang

Also mark API struct as initialized.

8 years agoEctor: Protect all headers by EFL_BETA_API_SUPPORT
Jean-Philippe Andre [Wed, 6 Jan 2016 05:26:25 +0000 (14:26 +0900)]
Ector: Protect all headers by EFL_BETA_API_SUPPORT

Those Ector APIs are not stable!

8 years agoevas: Update Ector_Color structure.
se.osadchy [Tue, 5 Jan 2016 11:41:12 +0000 (20:41 +0900)]
evas: Update Ector_Color structure.

Summary: Add a field at the end of the structure for defining the color encoding.

Reviewers: cedric, Hermet, raster, jpeg

Reviewed By: jpeg

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3530

8 years agoEvas 3d: Get hash table of scenes using the given node as root.
perepelits.m [Tue, 5 Jan 2016 11:39:02 +0000 (20:39 +0900)]
Evas 3d: Get hash table of scenes using the given node as root.

Summary: We need this to have an access to scene while running programs in edje-3d.

Reviewers: cedric, raster, Hermet, jpeg

Reviewed By: jpeg

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3529

8 years agoecore-wl2: Fix support for nested compositors
Chris Michael [Tue, 5 Jan 2016 18:10:40 +0000 (13:10 -0500)]
ecore-wl2: Fix support for nested compositors

When trying to create nested compositors, we were failing here due to
trying to read the WAYLAND_DISPLAY env variable and find an existing
server display to match. In order for nested compositors to function,
we cannot reuse that environment variable for server display hash.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-wl2: Remove blank line
Chris Michael [Tue, 5 Jan 2016 14:26:17 +0000 (09:26 -0500)]
ecore-wl2: Remove blank line

NB: No functional changes

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoeina_js: Fix documentation generation
Stefan Schmidt [Tue, 5 Jan 2016 16:09:31 +0000 (17:09 +0100)]
eina_js: Fix documentation generation

Summary:
Fix doxygen's documentation generation.
A bug in doxygen makes it fail with the _libv8_property_callback_info_test
class.
We put a #ifndef EFL_DOXYGEN around it to avoid it being seen by doxygen.

Add a NOTE comment about this.

ref T3005

@fix

Test Plan: make doc must succeed

Reviewers: felipealmeida, stefan_schmidt

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3534

8 years agoEctor gl: Ship gl eo headers dependend by the public header.
Tom Hacohen [Tue, 5 Jan 2016 15:56:22 +0000 (15:56 +0000)]
Ector gl: Ship gl eo headers dependend by the public header.

8 years agoEdje entry: Do not update cursor for non editable entry.
Subodh Kumar [Tue, 5 Jan 2016 14:30:20 +0000 (14:30 +0000)]
Edje entry: Do not update cursor for non editable entry.

Summary:
Do not update cursor for non editable entry.

If entry is not editable there must not be any need to query tb cursor
geometry and run other codes related to it. This may improve some
performance of non-editable entry.

Test Plan: NA

Reviewers: tasn, herdsman, thiepha

Subscribers: shilpasingh, cedric, jpeg

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3532

8 years agorelease: Update NEWS and bump version for 1.17.0-alpha1 release
Stefan Schmidt [Tue, 5 Jan 2016 13:57:42 +0000 (14:57 +0100)]
release: Update NEWS and bump version for 1.17.0-alpha1 release

8 years agoedje: Fix compiler warning about set but not used variables
Chris Michael [Tue, 5 Jan 2016 13:47:54 +0000 (08:47 -0500)]
edje: Fix compiler warning about set but not used variables

When compiling src/bin/edje, we end up with compiler warnings about
'pc' variable being set but not used. This patch just comments out
(for now) those variables (as I am unsure if they are going to be used
later or not).

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
8 years agopc: git ignore newly generated efl-js.pc file
Stefan Schmidt [Tue, 5 Jan 2016 09:37:00 +0000 (10:37 +0100)]
pc: git ignore newly generated efl-js.pc file

8 years agoEvas: (try to) fix compilation
Jean-Philippe Andre [Tue, 5 Jan 2016 09:59:22 +0000 (18:59 +0900)]
Evas: (try to) fix compilation

Not even sure if it's parallel compilation or not... classic case
of works-for-me.

Hopefully this should fix the build for others.

8 years agoEvas: Fix compilation for GLX
Jean-Philippe Andre [Tue, 5 Jan 2016 09:46:21 +0000 (18:46 +0900)]
Evas: Fix compilation for GLX

Sorry, I forgot to check that :)

8 years agowaylang-egl: Remove trailing spaces
Jean-Philippe Andre [Tue, 5 Jan 2016 08:06:18 +0000 (17:06 +0900)]
waylang-egl: Remove trailing spaces

8 years agowayland: Add support for TBM buffers
Jean-Philippe Andre [Tue, 5 Jan 2016 07:55:59 +0000 (16:55 +0900)]
wayland: Add support for TBM buffers

This is a follow up on the previous commits.
The original patch on Phab was split in 3 parts (factorization,
fix & this one).

Adds support to both SHM and EGL.

See D3501.

8 years agoevas tbm: Fix compilation warnings
Jean-Philippe Andre [Tue, 5 Jan 2016 07:54:48 +0000 (16:54 +0900)]
evas tbm: Fix compilation warnings

Previous patch used a wrong type... harmless but warned.

8 years agoEvas_Engine: add TBM surface and clean up Native Struct
Wonsik Jung [Tue, 5 Jan 2016 07:41:14 +0000 (16:41 +0900)]
Evas_Engine: add TBM surface and clean up Native Struct

Summary:
Add TBM surface in wayland and clean up Native struct on Evas Engine.
Previous TBM surface for evas_object_image's native_surface_set is only in Evas X11 backend.
This patch has the code for wayland backend.
In addition, evas_native_tbm.c is moved to software_generic. Becuase this file is common.
And, Native in Evas_Engine is clean-up.

@feature

Test Plan:
TBM surface is tested with spacegrapher's test app(tbm.c) in Tizen Device.
Pixmap surface is tested in ubuntu with same test app.
EvasGL is tested with elementary_test.

Reviewers: jpeg, spacegrapher, raster, cedric

Subscribers: dkdk, scholb.kim, JoogabYun

Differential Revision: https://phab.enlightenment.org/D3501

8 years agoEvas filters: Final fix after the ector refactor
Jean-Philippe Andre [Mon, 4 Jan 2016 12:46:56 +0000 (21:46 +0900)]
Evas filters: Final fix after the ector refactor

This implements a generic way of scaling buffers, using fake
RGBA_Image wrapping ector buffer maps. The underlying algo is
still the good old linear sw scaler.

Now the filters *should* be back to their previous level of
usability. Performance will probably be even worse than it was
before, for GL, as more glReadPixels may be involved. Optimization
now consists in actually implementing the filters with GL shaders.

8 years agoEvas filters: Complete basic repairs for GL engine
Jean-Philippe Andre [Mon, 4 Jan 2016 11:33:37 +0000 (20:33 +0900)]
Evas filters: Complete basic repairs for GL engine

Now the filters should work with the GL engine, again, but with
a potentially crazy performance. Indeed, the input buffer is now
backed by an FBO, that needs to be glReadPixel'ed everytime it is
accessed by the filters (mapped).

8 years agoEvas filters: Move evas image creation to the engine
Jean-Philippe Andre [Mon, 4 Jan 2016 09:00:30 +0000 (18:00 +0900)]
Evas filters: Move evas image creation to the engine

Evas filters is now more and more unaware of evas images (RGBA_Image)
and the engine + ector take care of everything.

Still left to do:
- map / unmap an FBO buffer into RO or RW memory

8 years agoEvas filters: Add GL buffer backed by RGBA_Image
Jean-Philippe Andre [Mon, 4 Jan 2016 08:48:48 +0000 (17:48 +0900)]
Evas filters: Add GL buffer backed by RGBA_Image

Dumb implementation of a "smart" buffer capable of wrapping an
RGBA_Image but that can still be rendered on screen (ie, an
Evas_GL_Image is attached to it).

8 years agoEvas: Fix make check after ector gl work
Jean-Philippe Andre [Mon, 4 Jan 2016 03:21:03 +0000 (12:21 +0900)]
Evas: Fix make check after ector gl work

evas_ector is not meant to be public so remove it from
the C++ bindings

8 years agoEvas filters & Ector GL: Prepare ground work for GL buffers
Jean-Philippe Andre [Thu, 31 Dec 2015 05:33:03 +0000 (14:33 +0900)]
Evas filters & Ector GL: Prepare ground work for GL buffers

This fixes crashes, adds safety, and notes a couple of things that
are not yet implemented:
- Make an Evas_GL_Image from an RGBA_Image so we can draw it on the
  canvas. This means Evas.Ector.GL.RGBA_Image.Buffer
- Readable Evas_GL_Image objects with gl_read_pixels
  --> Implement proper map() & unmap() for GL buffers

8 years agoEctor GL: Add skeletton for Evas.Ector.GL.Image.Buffer
Jean-Philippe Andre [Thu, 17 Dec 2015 09:04:06 +0000 (18:04 +0900)]
Ector GL: Add skeletton for Evas.Ector.GL.Image.Buffer

This is an ector buffer backed by an existing Evas_GL_Image

8 years agoEvas filters: Relax limitations about colorspaces
Jean-Philippe Andre [Fri, 11 Dec 2015 06:44:48 +0000 (15:44 +0900)]
Evas filters: Relax limitations about colorspaces

Since Ector Buffer implicitly converts colorspaces, we
can allow more commands to work even if they are suboptimal.

Now all filters should support any combinaison of input, map/mask
and output colorspaces.

8 years agoEvas filters: Fix proxies following previous refactor
Jean-Philippe Andre [Thu, 10 Dec 2015 06:50:40 +0000 (15:50 +0900)]
Evas filters: Fix proxies following previous refactor

FIXME:
 - GL support is still shaky

8 years agoEvas filters: Use Ector.Buffer instead of RGBA_Image
Jean-Philippe Andre [Wed, 9 Dec 2015 01:45:53 +0000 (10:45 +0900)]
Evas filters: Use Ector.Buffer instead of RGBA_Image

This is a major refactoring of the evas filters submodule.

Use Ector.Buffer and the map/unmap methods instead of directly
accessing image buffers with RGBA_Image. RGBA_Image is still
used under the hood, for two reasons:
 - Required for the final output (blend onto Evas itself)
 - Required for the scaling routines

FIXME:
 - Breaks proxy support (ie. all kind of texturing).
 - This breaks filters support for the GL engine.

8 years agoEvas filters: Minor debug improvement
Jean-Philippe Andre [Thu, 10 Dec 2015 06:12:23 +0000 (15:12 +0900)]
Evas filters: Minor debug improvement

8 years agoevas: add a way to tests Ector_GL.
Cedric BAIL [Mon, 4 Jan 2016 00:10:53 +0000 (01:10 +0100)]
evas: add a way to tests Ector_GL.

This is completely not supported/working/useful yet. Please do not
complain about it ruinning your life.

8 years agoector: fix typo during computation of bounds_get.
Cedric BAIL [Sun, 3 Jan 2016 21:04:23 +0000 (22:04 +0100)]
ector: fix typo during computation of bounds_get.

8 years agoector: add Ector.GL.Buffer.* support.
Jean-Philippe Andre [Thu, 17 Dec 2015 00:44:57 +0000 (16:44 -0800)]
ector: add Ector.GL.Buffer.* support.

8 years agoector: initial implementation of our own Ector GL backend.
Cedric BAIL [Thu, 12 Nov 2015 23:52:21 +0000 (15:52 -0800)]
ector: initial implementation of our own Ector GL backend.

8 years agoector: add drawable flag to buffers
Jean-Philippe Andre [Tue, 15 Dec 2015 10:58:43 +0000 (19:58 +0900)]
ector: add drawable flag to buffers

This indicates that a buffer can be used as a source to draw pixels.
Can't they all do that? Well, not exactly. A CPU buffer can't be drawn
by the GPU... not directly at least. That's what this flag is for.

8 years agoector eo: fix indentation (tabs --> spaces)
Jean-Philippe Andre [Mon, 14 Dec 2015 06:52:02 +0000 (15:52 +0900)]
ector eo: fix indentation (tabs --> spaces)

Why? Why would an editor replace 8 spaces by a TAB?
WHYYYYYY?

8 years agoector buffer: add COW access mode
Jean-Philippe Andre [Fri, 11 Dec 2015 09:49:16 +0000 (18:49 +0900)]
ector buffer: add COW access mode

In case you map a buffer once for read-only and once for write,
we can generate a temporary copy and return that instead. This
buffer will be copied back to the original surface once the COW
surface is unmapped.

8 years agoector: on-the-fly convert sw buffers during map()
Jean-Philippe Andre [Fri, 11 Dec 2015 06:23:55 +0000 (15:23 +0900)]
ector: on-the-fly convert sw buffers during map()

Also use map to generate spans.
This should simplify some filters code, making things work,
albeit inefficiently. At least they should work.

Fix doc too.

8 years agoector: track maps in software buffer
Jean-Philippe Andre [Thu, 10 Dec 2015 08:15:25 +0000 (17:15 +0900)]
ector: track maps in software buffer

8 years agoector: add engine-specific evas image buffer wrapper
Jean-Philippe Andre [Tue, 8 Dec 2015 08:29:54 +0000 (17:29 +0900)]
ector: add engine-specific evas image buffer wrapper

Since Evas still relies entirely on Image_Entry and Evas_GL_Image,
we will need an engine-specific wrapper object creating a Buffer
around an existing cached image.

Currently only SW support is implemented. GL will be more fun to
do (with glReadPixels and whatnot).

8 years agoector: add typedef Ector_Buffer
Jean-Philippe Andre [Tue, 8 Dec 2015 06:11:57 +0000 (15:11 +0900)]
ector: add typedef Ector_Buffer

8 years agoector: remove offset from Ector.Buffer.{map,unmap}
Jean-Philippe Andre [Wed, 9 Dec 2015 10:00:54 +0000 (19:00 +0900)]
ector: remove offset from Ector.Buffer.{map,unmap}

It just makes things a bit more complicated and doesn't correspond
to a classic "map" operation anyways.

Also return void* instead of uint8_t*. This is more correct and
avoid extra casts.

8 years agoevas: provide ector with a way to get GL symbol.
Cedric BAIL [Wed, 16 Dec 2015 23:19:57 +0000 (15:19 -0800)]
evas: provide ector with a way to get GL symbol.

8 years agoector: add an abstraction to get GL function.
Cedric BAIL [Wed, 16 Dec 2015 23:18:23 +0000 (15:18 -0800)]
ector: add an abstraction to get GL function.

8 years agoEvas filters: Fix crazy code (only when DEBUG=1)
Jean-Philippe Andre [Mon, 4 Jan 2016 05:20:17 +0000 (14:20 +0900)]
Evas filters: Fix crazy code (only when DEBUG=1)

8 years agoEo: Print object class name during Eo.Base ctor/dtor
Jean-Philippe Andre [Mon, 4 Jan 2016 08:44:02 +0000 (17:44 +0900)]
Eo: Print object class name during Eo.Base ctor/dtor

Printing class_name(MY_CLASS) is useless information.
Printing the object's real class name is much more useful.

8 years agoEvas: Replace abort() by ERR
Jean-Philippe Andre [Mon, 4 Jan 2016 12:29:57 +0000 (21:29 +0900)]
Evas: Replace abort() by ERR

8 years agoedje: custom state - if RTL flag is not set, return original description
Amitesh Singh [Tue, 5 Jan 2016 01:26:57 +0000 (06:56 +0530)]
edje: custom state - if RTL flag is not set, return original description

@fix

8 years agoeina: update example for with eina_base64url_decode function.
Srivardhan Hebbar [Mon, 4 Jan 2016 23:20:48 +0000 (15:20 -0800)]
eina: update example for with eina_base64url_decode function.

Summary:
Depends on D3521

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

Differential Revision: https://phab.enlightenment.org/D3523

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: add test case for eina_base64url_decode API.
Srivardhan Hebbar [Mon, 4 Jan 2016 23:20:04 +0000 (15:20 -0800)]
eina: add test case for eina_base64url_decode API.

Summary:
Depends on D3521

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

Differential Revision: https://phab.enlightenment.org/D3522

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: update .h file to make decodeurl api visible.
Srivardhan Hebbar [Mon, 4 Jan 2016 23:19:34 +0000 (15:19 -0800)]
eina: update .h file to make decodeurl api visible.

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

Reviewers: jpeg, cedric

Reviewed By: cedric

Differential Revision: https://phab.enlightenment.org/D3521

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoedje_cc: abort when insert_before/after part doesn't exist
Jee-Yong Um [Mon, 4 Jan 2016 23:15:31 +0000 (15:15 -0800)]
edje_cc: abort when insert_before/after part doesn't exist

Summary:
edje_cc does not check whether insert_before/after part exists.
edje_cc will notify a compilation error to developer when insert_before
or insert_after part does not exist.

T2513

Reviewers: Hermet, NikaWhite

Subscribers: cedric, jpeg

Maniphest Tasks: T2513

Differential Revision: https://phab.enlightenment.org/D3517

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: return NULL iterator in all case, not only when safety check is turned on.
Cedric BAIL [Mon, 4 Jan 2016 23:13:06 +0000 (15:13 -0800)]
eina: return NULL iterator in all case, not only when safety check is turned on.

8 years agoedje: add edje_object_color_class_clear()
Jee-Yong Um [Mon, 4 Jan 2016 23:10:58 +0000 (15:10 -0800)]
edje: add edje_object_color_class_clear()

Summary: clear all color classes defined in edje object

Reviewers: raster, jpeg

Subscribers: woohyun, cedric, kimcinoo

Differential Revision: https://phab.enlightenment.org/D3514

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: create Eina_Vector2 structure and add functions.
se.osadchy [Mon, 4 Jan 2016 23:05:58 +0000 (15:05 -0800)]
evas: create Eina_Vector2 structure and add functions.

Summary: Only copy code from evas_3d_utils, rename evas to eina and add documentation.

Reviewers: Hermet, raster, jpeg, cedric

Reviewed By: jpeg, cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3509

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: fix convex hull.
perepelits.m [Mon, 4 Jan 2016 22:50:42 +0000 (14:50 -0800)]
evas: fix convex hull.

Summary:
There were some problems with second and third vertices in the first triangle of convex hull. It is very hard to see this errors because it could cause
an excess of triangles inside of convex hull, I have used blender to find them.

Reviewers: raster, Hermet, cedric

Reviewed By: cedric

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3491

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoedje: implementation of Edje_3d
perepelits.m [Mon, 4 Jan 2016 22:03:12 +0000 (14:03 -0800)]
edje: implementation of Edje_3d

Summary:
I have a strange feeling that I always commit the same code to edje-3d, I hope it will finish very soon.
However here are some important new keywords, descriptors for them, methods for models,
structure for providing animation in edje programs and some more changes.

Reviewers: raster, Hermet, cedric

Reviewed By: cedric

Subscribers: jpeg, artem.popov

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3437

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_con: from ecore_con_connector to efl_network_connector.
Srivardhan Hebbar [Mon, 4 Jan 2016 20:35:37 +0000 (12:35 -0800)]
ecore_con: from ecore_con_connector to efl_network_connector.

Summary:
Changed ecore_con_connector.eo to efl_network_connector.eo as part of
migrating to efl_network.

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

Differential Revision: https://phab.enlightenment.org/D3427

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: update example with eina_strbuf_manage_read_only_new_length function.
Srivardhan Hebbar [Mon, 4 Jan 2016 20:32:15 +0000 (12:32 -0800)]
eina: update example with eina_strbuf_manage_read_only_new_length function.

Summary:
The example is to demonstrate the use of
eina_strbuf_manage_read_only_new_length API.

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

Differential Revision: https://phab.enlightenment.org/D3421

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: fix ply example after adding of new savers and loaders.
perepelits.m [Mon, 4 Jan 2016 20:31:20 +0000 (12:31 -0800)]
evas: fix ply example after adding of new savers and loaders.

Summary:
Before adding of new saver and loader for ply format we had a chance to set EVAS_CANVAS3D_SHADE_MODE_VERTEX_COLOR
for every mesh, now we save only existing atributes and there are a lot of meshes we have to set EVAS_CANVAS3D_SHADE_MODE_VERTEX_COLOR.

[FIX]

Reviewers: cedric, raster, Hermet

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3412

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: factorize of data structure with evas_color and ector.
se.osadchy [Mon, 4 Jan 2016 20:16:44 +0000 (12:16 -0800)]
evas: factorize of data structure with evas_color and ector.

Summary: Move data structure and functionality to ector_color from evas_color.

Reviewers: Hermet, raster, cedric

Reviewed By: cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3393

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore-wl2: Remove extra blank line
Chris Michael [Mon, 4 Jan 2016 16:05:57 +0000 (11:05 -0500)]
ecore-wl2: Remove extra blank line

NB: No functional changes

Signed-off-by: Chris Michael <cp.michael@samsung.com>
8 years agoEvas textblock: Clean up ellipsis handling code.
Tom Hacohen [Mon, 4 Jan 2016 15:09:16 +0000 (15:09 +0000)]
Evas textblock: Clean up ellipsis handling code.

This cleans up the previous commit a bit by isolating the calculation
code and using a temporary descriptively named variable for the calculation.

8 years agoEvas Textblock: Fix ellipsis when textblock is resized to formatted h.
Youngbok Shin [Mon, 4 Jan 2016 15:00:56 +0000 (15:00 +0000)]
Evas Textblock: Fix ellipsis when textblock is resized to formatted h.

Summary:
it->h is sum of max ascent and max descent. It shouldn't be used
when handle ellipsis. Because, Evas Textblock uses these values for
each lines differently according to its location.
(start, end, else, single)
So, for handling ellipsis exactly, it has to be fixed.

Test Plan: A test case is included in Evas Test suite.

Reviewers: woohyun, tasn, herdsman

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3475

8 years agoecore-drm: Add a new API for keymap cache
JengHyun Kang [Mon, 4 Jan 2016 13:47:43 +0000 (08:47 -0500)]
ecore-drm: Add a new API for keymap cache

Summary:
Originally, each keyboard devices could have their own keymap.
The one keyboard's keymap could different with others.
But for this, Ecore_Drm compile a new keymap when a keyboard is connected.
But this is a burden for some people who doesn't manage keymap for each keyboard.
They want to maintain only one keymap for system.
So, I added cached context/keymap and just ref/unref for each keyboard device.
People who want to different keymap for each keyboard just do not set cached
context/keymap. Then Ecore_Drm maintain keymaps about each keyboard devices.

Test Plan:
Connect a keyboard device and watch flow of ioctl.
Originally Ecore_Drm opened xkb data and compile keymap,
but after patch, that ioctl is disppeared.

@feature

Reviewers: raster, devilhorns, ManMower

Reviewed By: devilhorns, ManMower

Subscribers: cedric, input.hacker, ohduna, jpeg

Differential Revision: https://phab.enlightenment.org/D3503

8 years agoedje: enhance circular dependency error message of edje calculation
Shinwoo Kim [Thu, 31 Dec 2015 04:49:56 +0000 (13:49 +0900)]
edje: enhance circular dependency error message of edje calculation

Summary: enhance error message of edje calculation

Test Plan: check edje having circular dependency

Reviewers: raster, cedric, jpeg

Subscribers: seoz

Differential Revision: https://phab.enlightenment.org/D3484

8 years agoeina: Updated example program to demonstrate base64url encode.
Srivardhan Hebbar [Thu, 31 Dec 2015 04:34:19 +0000 (13:34 +0900)]
eina: Updated example program to demonstrate base64url encode.

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

Reviewers: cedric, jpeg

Reviewed By: jpeg

Differential Revision: https://phab.enlightenment.org/D3483

8 years agoedje example: change wrong comment, and remove warning message by adding ellipsis
Shinwoo Kim [Thu, 31 Dec 2015 02:59:57 +0000 (11:59 +0900)]
edje example: change wrong comment, and remove warning message by adding ellipsis

Summary: Enhance edje color class example

Test Plan: build using the comment message

Reviewers: raster, cedric, jpeg

Reviewed By: jpeg

Subscribers: seoz

Differential Revision: https://phab.enlightenment.org/D3508

8 years agoedje_embryo: add printf feature.
Jaehwan Kim [Thu, 31 Dec 2015 02:59:48 +0000 (11:59 +0900)]
edje_embryo: add printf feature.

Summary:
The edc debugging is difficult because users can not check
whether or not any program is executed or the state of part.
If users can print that property, they can do debugging more easily.
This "printf" feature is for debugging edc.

@feature

Test Plan:
add printf("test : %s %d %f", "text", 1, 0.5); in the script in the edc
build it by edje_cc
excute it and see the log

Reviewers: cedric, raster, jpeg

Reviewed By: jpeg

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3499

8 years agoEdje_cc: support braces for description.proxy attribute.
Mykyta Biliavskyi [Wed, 30 Dec 2015 07:03:30 +0000 (16:03 +0900)]
Edje_cc: support braces for description.proxy attribute.

Fix parsing Group.Parts.Part.Description.Proxy block in cases
when uses braces:

part { type: PROXY;
  description {
    ..
    proxy {
      source_clip:    1;
      source_visible: 1;
    }
      ..
  }
}

@fix

8 years agoembryo: add eina_log
Jaehwan Kim [Wed, 30 Dec 2015 00:14:36 +0000 (09:14 +0900)]
embryo: add eina_log

Summary: add eina_log

Test Plan: use DBG, ERR ... in embryo code

Reviewers: raster, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3507

8 years agoecore-drm: Add missing variable 'dev'
Chris Michael [Tue, 29 Dec 2015 14:12:50 +0000 (09:12 -0500)]
ecore-drm: Add missing variable 'dev'

Seems this should have been declared in the previously pushed patch,
but was missing.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoAdd a new Ecore_Drm API for setting left-handed mouse
JengHyun Kang [Tue, 29 Dec 2015 14:03:43 +0000 (09:03 -0500)]
Add a new Ecore_Drm API for setting left-handed mouse

Summary:
Add a new Ecore_Drm API named ecore_drm_device_left_handed_set().
Libinput support various input options. Left handed mode is one of options.
Right-handed people and left-handed people use a mouse
using different button mapping.
So if a left handed option is enabled, libinput change right mouse button and
left mouse button.
So support this option, I added this api using libinput's left handed option.

@feature

Test Plan:
After set left handed mode, left mouse button generate button numbered 3,
and right mouse button generate button numbered 1.

Reviewers: raster, ManMower, devilhorns

Subscribers: ohduna, cedric, jpeg, input.hacker

Differential Revision: https://phab.enlightenment.org/D3431

8 years agoeo-cxx: Removed pessimization use of std::move
Felipe Magno de Almeida [Tue, 29 Dec 2015 13:38:50 +0000 (11:38 -0200)]
eo-cxx: Removed pessimization use of std::move

Local variables, when returned, are already moved or copy-elised. The
use of std::move removes the possibility of copy-elision, making it
possibly slower. Removed the wrong std::move use.

8 years agoeldbus-cxx: Fixed infinite recursion with post-increment operator
Felipe Magno de Almeida [Tue, 29 Dec 2015 13:37:27 +0000 (11:37 -0200)]
eldbus-cxx: Fixed infinite recursion with post-increment operator

The pre-increment operator was being defined twice and being
implemented as a post-increment. Modified declaration to be
post-increment as it was supposed to be.

8 years agoeina-cxx: Fixed wrong uses of using directive
Felipe Magno de Almeida [Tue, 29 Dec 2015 13:36:11 +0000 (11:36 -0200)]
eina-cxx: Fixed wrong uses of using directive

Changed using directive for types to typedef's to bring their definitions from base type. Avoiding compilation errors with clang.

8 years agoEvil: fix compilation warning
Vincent Torri [Mon, 28 Dec 2015 22:15:29 +0000 (23:15 +0100)]
Evil: fix compilation warning

getenv() requires stdlib.h

8 years agoEvil: integrate the dlfcn code into Evil
Vincent Torri [Mon, 28 Dec 2015 13:22:29 +0000 (14:22 +0100)]
Evil: integrate the dlfcn code into Evil

This will remove some incompatibilities with other packages,
especially for win-builds

8 years agoEvas: Implement ARGY88 and GRY8 conversion to ARGB8888
Jean-Philippe Andre [Tue, 29 Dec 2015 13:01:16 +0000 (22:01 +0900)]
Evas: Implement ARGY88 and GRY8 conversion to ARGB8888

This is adding support for a deprecated function, but apparently
it is used.

Also, remove crazy abort() inside the data_put function.

Fixes T2991

8 years agoEo: Prevent shutdown from actually doing anything
Jean-Philippe Andre [Tue, 29 Dec 2015 11:29:36 +0000 (20:29 +0900)]
Eo: Prevent shutdown from actually doing anything

Currently, eo_shutdown can't work.

Every Eo_Class ID is stored inside its class_get() function as a
static variable. This means any call to class_get() after eo_shutdown()
(even if eo_init was done properly) will lead to using an invalid ref
for the class id. In other words, the class is not valid anymore,
and objects can't be created.

Resetting the pointer to NULL would be possible, if we passed it
during the class creation. But this would lead to potential crashes
if a class was created from a now dlclosed library.

The only solution I can envision here is to check that class_get
actually returns a valid ref with the right class name. Most likely
the performance impact is not acceptable.

This fixes make check for me (with systemd module for ecore).

8 years agoEo: Add internal function call on pointer errors
Jean-Philippe Andre [Tue, 29 Dec 2015 10:42:05 +0000 (19:42 +0900)]
Eo: Add internal function call on pointer errors

Useful for GDB: break on this function when things go wrong.
Similar to eina_safety.

I guess we could set some Eina_Error and maybe even have error
callbacks for easier application debugging. Later.

8 years agoEcore Eldbus: Work around circular dependencies
Jean-Philippe Andre [Tue, 29 Dec 2015 08:06:53 +0000 (17:06 +0900)]
Ecore Eldbus: Work around circular dependencies

eldbus initializes ecore that may then init eldbus again,
since one of the systemd modules is for eldbus.
eldbus_shutdown() is then no longer functional, as there are
two refs on eldbus.

This patch solves this problem by removing the extra ref on
eldbus from the module if it was already initialized.

This patch now introduces really bad issues since there are now
EO classes that fail to work after module load-unload-reload.

8 years agoEet: Fail gracefully on invalid input
Jean-Philippe Andre [Tue, 29 Dec 2015 06:56:24 +0000 (15:56 +0900)]
Eet: Fail gracefully on invalid input

Consider infinity and NaN as invalid input for
put_float and put_double, since the underlying convert
function can't process them.

This fixes potential errors with incomplete / invalid evas 3d models

8 years agoEina: Fix potential crash in eina_convert_dtoa
Jean-Philippe Andre [Tue, 29 Dec 2015 06:45:35 +0000 (15:45 +0900)]
Eina: Fix potential crash in eina_convert_dtoa

Infinity and NaN should probably just not be converted.

An error would happen in make check while saving evas 3d models.

8 years agoEvas 3d: Fix some potential crashes
Jean-Philippe Andre [Tue, 29 Dec 2015 06:41:51 +0000 (15:41 +0900)]
Evas 3d: Fix some potential crashes

Those were happening before the previous patches.

8 years agoEvas 3d: Fix some issues in model test case
Jean-Philippe Andre [Tue, 29 Dec 2015 06:02:52 +0000 (15:02 +0900)]
Evas 3d: Fix some issues in model test case

Fix some errors, check that file_set and file_save actually work
(rather than have them both fail and pass the test since 0 == 0),
fix temp paths and cleanup unused file.

8 years agoEvas 3d: Improve error messages in model load & save
Jean-Philippe Andre [Tue, 29 Dec 2015 05:51:28 +0000 (14:51 +0900)]
Evas 3d: Improve error messages in model load & save

See make check's evas_suite ouput

8 years agoEvas: Fix potential crash with extraneous eet_shutdown
Jean-Philippe Andre [Tue, 29 Dec 2015 05:33:21 +0000 (14:33 +0900)]
Evas: Fix potential crash with extraneous eet_shutdown

Shutdown eet only if it was actually initialized.

This fixes make check (for me) after applying the previous patch.
It seems on my machine, the gl cache files are not saved properly
from within the make check environment.

8 years ago[Evas] Refactor model's savers and loaders.
perepelits.m [Tue, 29 Dec 2015 04:28:55 +0000 (13:28 +0900)]
[Evas] Refactor model's savers and loaders.

Summary:
Move common part to a separated document.
Make code more readable using smaller functions. (from Task T2713)

Everything is OK with make check.

Reviewers: cedric, raster, Hermet, stefan_schmidt

Reviewed By: stefan_schmidt

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3430

8 years agoecore_con: fix typo preventing IPv6 multicast to work.
Cedric BAIL [Mon, 28 Dec 2015 22:01:15 +0000 (14:01 -0800)]
ecore_con: fix typo preventing IPv6 multicast to work.

Thanks Vincent Torri for spotting it.

8 years agoeo: Fix resource leak
Chris Michael [Mon, 28 Dec 2015 15:09:36 +0000 (10:09 -0500)]
eo: Fix resource leak

Coverity CID1339783 says that we have a potential resource leak here.
'cb' gets allocated via calloc, but is not freed if we end up
returning here

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-drm: Add APIs to support key remap functionality
Sung-Jin Park [Mon, 28 Dec 2015 14:18:05 +0000 (09:18 -0500)]
ecore-drm: Add APIs to support key remap functionality

Summary:
This adds two new APIs to enable/set key remap functionality and
a number of keys to be remapped to the other keys. As of now there is no
api to do this therefore we need to remap using linux utility such as
'setkeycodes'. By adding/calling these apis, each Ecore_Drm_Evdev device
will have its specific key remap hash and we can apply each remapping keys
for each key/keyboard device.

Test Plan:
(1) Enable key remap and set remapping of a key on a specific keyboard device
(2) Plug in the keyboard device and check the key is being remapped or not
(3) Check the other keys are coming normally
(4) Check the the remapping key on a specific keyboard doesn't affect to any other devices

Signed-off-by: Sung-Jin Park <input.hacker@gmail.com>
Reviewers: raster, zmike, gwanglim, ManMower, devilhorns

Subscribers: JHyun, ohduna, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3463

8 years agoecore-wl2: Fix MIN() macro comparing int vs unsigned int
Chris Michael [Mon, 28 Dec 2015 14:15:01 +0000 (09:15 -0500)]
ecore-wl2: Fix MIN() macro comparing int vs unsigned int

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>