platform/upstream/efl.git
9 years agoeldbus hackaround for proxy deletion during pending callback
Mike Blumenkrantz [Mon, 8 Dec 2014 19:55:30 +0000 (14:55 -0500)]
eldbus hackaround for proxy deletion during pending callback

this is merely a way to avoid crashing immediately when the referenced bug is triggered. it does not fix the problem.

ref T1908

9 years agoelua: add pc file into gitignore
Daniel Kolesa [Mon, 8 Dec 2014 16:13:02 +0000 (16:13 +0000)]
elua: add pc file into gitignore

9 years agoelua: add elua.pc to configure.ac and add cmake definitions for elua
Daniel Kolesa [Mon, 8 Dec 2014 16:11:35 +0000 (16:11 +0000)]
elua: add elua.pc to configure.ac and add cmake definitions for elua

9 years agoelua: add pkg-config for the elua library
Daniel Kolesa [Mon, 8 Dec 2014 15:59:13 +0000 (15:59 +0000)]
elua: add pkg-config for the elua library

9 years agoeina: Refactored code and removed unwanted assignment.
Srivardhan Hebbar [Mon, 8 Dec 2014 13:46:50 +0000 (14:46 +0100)]
eina: Refactored code and removed unwanted assignment.

Summary:
Removed unused assignment for length.

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

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1739

9 years agoelua: initial skeleton for the Elua library
Daniel Kolesa [Mon, 8 Dec 2014 11:38:04 +0000 (11:38 +0000)]
elua: initial skeleton for the Elua library

Because of Bob and other things, it is required to separate the functionality currently
provided by the elua binary into a library. This library will then be used by the elua
binary as well as any other project.

9 years agoevas/gl_drm: Fix warning
Gwanglim Lee [Mon, 8 Dec 2014 10:07:27 +0000 (19:07 +0900)]
evas/gl_drm: Fix warning

Using int instead of uint32_t.

9 years agoEvas font: Compress font algo should round instead of floor
Jean-Philippe Andre [Mon, 8 Dec 2014 04:50:44 +0000 (13:50 +0900)]
Evas font: Compress font algo should round instead of floor

This is an attempt at having higher quality font rendering
while still using RLE-based font compression.

9 years agoeina_thread: fix memory leak
kabeer khan [Mon, 8 Dec 2014 08:19:28 +0000 (09:19 +0100)]
eina_thread: fix memory leak

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

Reviewers: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1737

9 years agoeina: corrected documentation of ecore_hash_destroy function in benchmarj.
Srivardhan Hebbar [Mon, 8 Dec 2014 08:18:08 +0000 (09:18 +0100)]
eina: corrected documentation of ecore_hash_destroy function in benchmarj.

Summary:
ecore_hash_destroy function return nothing, still in the documentation of the function, it was mentioned that it would return true or false. So corrected the documentation.

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

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1738

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoecore-wayland: make possible to disable xdg-shell at runtime
Philippe Coval [Fri, 5 Dec 2014 19:02:16 +0000 (14:02 -0500)]
ecore-wayland: make possible to disable xdg-shell at runtime

Summary:
Since xdg-shell is not yet supported officialy by wayland
It makes sense to have it disabled by default
and enable it on purpose.

Using env variable : EFL_WAYLAND_DONT_USE_XDG_SHELL
(name is inspired from Qt implementation of xdg-shell
but in the opposite way)

There is no EFL_WAYLAND_USE_XDG_SHELL in efl,
please use EFL_WAYLAND_DONT_USE_XDG_SHELL instead.

By default xdg-shell is enabled since it's supported by efl

Since there is only one shell supported at runtime
it's easier to switch to wl-shell (wayland's fallback shell)
by change this env variable.

Note, this patch can be reverted once xdg-shell replaces wl-shell
in wayland which is not the case in weston-1.6
but could integrated into upcoming wayland-1.7
(to be confirmed on release)

Change-Id: Id3732492397df9abe4a7c9e6e92a8f2c993c8395
Bug: https://phab.enlightenment.org/T1901
Bug-Tizen: TC-1353/part
Forwarded: https://phab.enlightenment.org/T1901
Signed-off-by: Philippe Coval <philippe.coval@open.eurogiciel.org>
Tizen Test Plan: echo 'export EFL_WAYLAND_DONT_USE_XDG_SHELL=defined' > /etc/profile.d/ecore.sh

Reviewers: seoz, devilhorns

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1734

9 years agoeeze_scanner now has separate return codes for various startup-related failures
Mike Blumenkrantz [Fri, 5 Dec 2014 18:35:14 +0000 (13:35 -0500)]
eeze_scanner now has separate return codes for various startup-related failures

ref T1612

9 years agoeeze_scanner needs to hash on the address of the pointer and not use direct_add()
Mike Blumenkrantz [Fri, 5 Dec 2014 18:34:11 +0000 (13:34 -0500)]
eeze_scanner needs to hash on the address of the pointer and not use direct_add()

this is an @fix for the constant crashing

ref T1612

9 years agoecore-con now prints error messages when local socket creation fails
Mike Blumenkrantz [Fri, 5 Dec 2014 18:19:49 +0000 (13:19 -0500)]
ecore-con now prints error messages when local socket creation fails

9 years agoecore-evas-wayland: Fix issue with ELM_PROFILE=mobile and windows not
Chris Michael [Fri, 5 Dec 2014 17:49:42 +0000 (12:49 -0500)]
ecore-evas-wayland: Fix issue with ELM_PROFILE=mobile and windows not
showing up correctly.

Summary: If elm config profile is set to Mobile, then the
auto_norender_withdrawn is set to 1, which was causing Elm windows to
not draw properly due to the ee->withdrawn property not being updated.
It was not getting updated due to these checks for override.

NB: Tested this with all elm profiles, tested in weston, and tested in
Enlightenment (Wayland).

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeldbus: Use the timeout parameter on eldbus_proxy_send_and_block()
José Roberto de Souza [Fri, 5 Dec 2014 03:04:20 +0000 (01:04 -0200)]
eldbus: Use the timeout parameter on eldbus_proxy_send_and_block()

Sorry by bad I had forgot to use the new parameter.

9 years agoeldbus: Add timeout parameter to eldbus_proxy_send_and_block()
José Roberto de Souza [Fri, 5 Dec 2014 03:04:20 +0000 (01:04 -0200)]
eldbus: Add timeout parameter to eldbus_proxy_send_and_block()

As all other send functions have, we need to have here.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeldbus: Fix _eldbus_connection_send_and_block()
José Roberto de Souza [Fri, 5 Dec 2014 03:01:37 +0000 (01:01 -0200)]
eldbus: Fix _eldbus_connection_send_and_block()

Some problems with the actual implementation:
- the reply should not be writable, as it can only be read.
- if an error happen dbus_connection_send_with_reply_and_block()
will return NULL so we need check before use it
- all other send calls remove one reference of the message

Now also it is creating a error message, so the caller can know why it fail.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore/animator: simpler code
ChunEon Park [Fri, 5 Dec 2014 14:24:16 +0000 (23:24 +0900)]
ecore/animator: simpler code

9 years agoecore/animator: fix the bezier cubic compuation.
ChunEon Park [Fri, 5 Dec 2014 14:17:38 +0000 (23:17 +0900)]
ecore/animator: fix the bezier cubic compuation.

The previous compuation is totally wrong.
Even it doesn't work correctly.
(I have no idea what the orignal author was thinking?)

Here we just need a simple and clear fomular to get the current progress frame.

If i'm wrong, please ping me.

@fix

9 years agoEo: fix error handling when too many deletions invocations occur.
Daniel Zaoui [Fri, 5 Dec 2014 01:19:47 +0000 (03:19 +0200)]
Eo: fix error handling when too many deletions invocations occur.

Before this fix, when a deletion was invoked twice on an object, a
wrong message ("...You wrongly call eo_unref() within a destructor...")
was printed. This was caused by the del_triggered flag that was not
resetted when the destruction finished.

This patch fixes this behavior by printing the right message on a double
deletion.

9 years agoeldbus: Add actual API function code for eldbus_proxy_send_and_block
Chris Michael [Thu, 4 Dec 2014 14:48:21 +0000 (09:48 -0500)]
eldbus: Add actual API function code for eldbus_proxy_send_and_block

Summary: This adds the actual code to send a dbus message and block
while waiting for a reply.

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeldbus: Add code for _eldbus_connection_send_and_block function
Chris Michael [Thu, 4 Dec 2014 14:47:01 +0000 (09:47 -0500)]
eldbus: Add code for _eldbus_connection_send_and_block function

Summary: This function will send a message to dbus and block while
waiting for a reply

NB: This is needed for our 'port to libinput', and for our 'opening up the
drm card without systemd' efforts

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeldbus: Add API function declaration for eldbus_proxy_send_and_block
Chris Michael [Thu, 4 Dec 2014 14:45:51 +0000 (09:45 -0500)]
eldbus: Add API function declaration for eldbus_proxy_send_and_block

Summary: This adds a public facing API function to make dbus calls
which will block and wait for a reply. This is needed for a couple of
use cases in our Wayland efforts (libinput, etc).

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoeldbus: Add function prototype for send_and_block
Chris Michael [Thu, 4 Dec 2014 14:43:50 +0000 (09:43 -0500)]
eldbus: Add function prototype for send_and_block

Summary: This just adds the function prototype into the eldbus private
header. It will be used in the new proxy function
"eldbus_proxy_send_and_block"

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore: Fix @since version to 1.13 for new API
Stefan Schmidt [Thu, 4 Dec 2014 14:02:34 +0000 (15:02 +0100)]
ecore: Fix @since version to 1.13 for new API

1.3 is acient so we really will release this with 1.13.
Removed some trailing whitespaces devilhorns added. :)

9 years agoecore: Add code for new API function 'ecore_main_loop_nested_get'
Chris Michael [Thu, 4 Dec 2014 13:56:43 +0000 (08:56 -0500)]
ecore: Add code for new API function 'ecore_main_loop_nested_get'

Summary: This commit adds the actual code to the function, which
returns the 'in_main_loop' variable so that we can detect if the
ecore_main_loop is actually running.

NB: Will be needed for new eldbus API function (yet to add).

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore: Add a new API function ecore_main_loop_nested_get
Chris Michael [Thu, 4 Dec 2014 13:54:48 +0000 (08:54 -0500)]
ecore: Add a new API function ecore_main_loop_nested_get

Summary: This commit adds a new function 'ecore_main_loop_nested_get'
so that we can detect if the ecore_main_loop is running.

NB: This is going to be needed for a new eldbus function that we have
to add in order to handle a use-case on the Wayland side. Spoke with
cedric for a while wrt to all this, and he gave it his 'ok' ;)

@feature

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoconfigure: Add note about automake 1.14 problem
Stefan Schmidt [Thu, 4 Dec 2014 13:11:35 +0000 (14:11 +0100)]
configure: Add note about automake 1.14 problem

Maybe we can remove this once it is fixed and the fixed version it widely used.

9 years agobuild: support automake-1.14
Philippe Coval [Thu, 4 Dec 2014 12:50:58 +0000 (13:50 +0100)]
build: support automake-1.14

Summary:
This was needed to install example on purposes

The problem is that the install_sh* variables
in generated makefiles are expanding to ~/install-sh
which is wrong and should be the one in root source tree

Change-Id: I5311298864a5f042de90071cb6dfaefbefdf001a
Bug: https://phab.enlightenment.org/T1889
Bug-Tizen: TC-2177
Signed-off-by: Philippe Coval <philippe.coval@open.eurogiciel.org>
Reviewers: seoz

Subscribers: stefan_schmidt, cedric

Maniphest Tasks: T1889

Differential Revision: https://phab.enlightenment.org/D1724

9 years agoevas/examples: fix warnings.
ChunEon Park [Thu, 4 Dec 2014 10:47:41 +0000 (19:47 +0900)]
evas/examples: fix warnings.

removed unused.

9 years agoevas/examples: fix build break
ChunEon Park [Thu, 4 Dec 2014 10:42:27 +0000 (19:42 +0900)]
evas/examples: fix build break

add dependency libefl to call efl_file_set().

9 years agoevas/gl_drm: add native surface for wayland_egl_client.
Mun, Gwan-gyeong [Thu, 4 Dec 2014 08:59:30 +0000 (17:59 +0900)]
evas/gl_drm: add native surface for wayland_egl_client.

Reviewers: zmike, devilhorns, cedric, gwanglim

Reviewed By: cedric, gwanglim

Subscribers: jpeg, torori, cedric

Differential Revision: https://phab.enlightenment.org/D1507

9 years agoevas - gif loader - fix technically "wrong" sizeof (doesnt cause a bug)
Carsten Haitzler (Rasterman) [Thu, 4 Dec 2014 07:03:32 +0000 (16:03 +0900)]
evas - gif loader - fix technically "wrong" sizeof (doesnt cause a bug)

go from sizeof(char **) to sizeof(char *) ... effectively. so no real bug.

9 years agoecore-drm: Remove erroneous debug messages
Chris Michael [Wed, 3 Dec 2014 19:18:52 +0000 (14:18 -0500)]
ecore-drm: Remove erroneous debug messages

Summary: No real function changes, just some cleanup of old debug
messages

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoelua: support is_c_only in lualian
Daniel Kolesa [Wed, 3 Dec 2014 15:27:38 +0000 (15:27 +0000)]
elua: support is_c_only in lualian

9 years agoeolian: eolian_function_is_c_only tests
Daniel Kolesa [Wed, 3 Dec 2014 15:04:40 +0000 (15:04 +0000)]
eolian: eolian_function_is_c_only tests

9 years agoeolian: add eolian_function_is_c_only
Daniel Kolesa [Wed, 3 Dec 2014 14:59:24 +0000 (14:59 +0000)]
eolian: add eolian_function_is_c_only

This function allows us to mark functions that are not bindable.
Also remove some obsolete code.

9 years agoecore con: Fix GnuTLS build error.
Andreas Metzler [Wed, 3 Dec 2014 12:07:40 +0000 (12:07 +0000)]
ecore con: Fix GnuTLS build error.

By making the respective changes on the GnuTLS side for upstream commits
d9b5f192d4883193f79cd3e43ed1da52521825dc and
d72f809fb874c8f14b1461949802905d41d648a1
which only take care of OpenSSL.

@fix

9 years agoedje/edje_cc_handler - updated internal doc.
ChunEon Park [Wed, 3 Dec 2014 12:09:06 +0000 (21:09 +0900)]
edje/edje_cc_handler - updated internal doc.

9 years agoEvas: Evas_3D - Add check of visibilty node.
Dmytro Dadyka [Wed, 3 Dec 2014 08:48:55 +0000 (09:48 +0100)]
Evas: Evas_3D - Add check of visibilty node.

If the node is not visible, it is not rendered, which improves performance.
@feature.

Reviewers: raster, Hermet, cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1722

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoautotools: fix compiling/linking gl engines on OSX.
Jean Guyomarc'h [Wed, 3 Dec 2014 08:45:51 +0000 (09:45 +0100)]
autotools: fix compiling/linking gl engines on OSX.

Summary:
Until now, it was necessary to set global LDFLAGS and CFLAGS to allow
compiling (and linking) engines using OpenGL.
gl_generic used to complained about missing headers or unkown libraries.
A problem on OSX is that there is CGL (Apple's Core OpenGL) on which the whole system
relies on and GLX, when X11 is installed; and they cohabit together.
When gl_cocoa is enabled, gl_generic is now linked against CGL.
When it is not, gl_generic is compiled with and linked against GLX as a fallback.

@fix

Test Plan:
With --enable-cocoa: software_x11, opengl_x11 and opengl_cocoa are working as expected.
With --disable-cocoa: software_x11 and opengl_x11 are also working as expected.
No compiling nor linking problems have been issued.

Reviewers: cedric, raster, raoulh

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1723

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: Evas_3D - fix typo.
perepelits.m [Wed, 3 Dec 2014 08:42:50 +0000 (09:42 +0100)]
evas: Evas_3D - fix typo.

Summary:
A little fix of copy-paste, there were problems while changing texture coordinates of indices.

@fix

Reviewers: raster, Hermet, cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1725

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoEvas GL: Clarify doc about Evas_GL_API struct version
Jean-Philippe Andre [Wed, 3 Dec 2014 03:08:03 +0000 (12:08 +0900)]
Evas GL: Clarify doc about Evas_GL_API struct version

9 years agoEvas GL: Add preventive padding after Evas_GL_API
Jean-Philippe Andre [Wed, 3 Dec 2014 02:44:09 +0000 (11:44 +0900)]
Evas GL: Add preventive padding after Evas_GL_API

Since this struct is likely to grow in size over time, client apps
built against future versions of EFL might start indexing fields
that are not present in the current form.

Also, don't reset the struct memory as this would break
multithreaded GL applications.

While this is not exactly a fix, I'll backport this.

@fix

9 years agoevas-gl-drm: remove initalizing gbm on engine setup
MinJeong Kim [Wed, 3 Dec 2014 07:30:13 +0000 (16:30 +0900)]
evas-gl-drm: remove initalizing gbm on engine setup

Summary: Removed evas_drm_gbm_init() on the initial setup of evas gl_drm engine.

Reviewers: gwanglim

Reviewed By: gwanglim

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1728

9 years agoEvas GL: Update evas gl api version number to 3
Gwanglim Lee [Wed, 3 Dec 2014 02:48:03 +0000 (11:48 +0900)]
Evas GL: Update evas gl api version number to 3

evas gl api supports wayland extensions as being part of a struct
of version 3.

9 years agoEvas GL: Add support for bind_wayland_display extension
Gwanglim Lee [Tue, 2 Dec 2014 13:45:36 +0000 (22:45 +0900)]
Evas GL: Add support for bind_wayland_display extension

This provides EGL_WL_bind_wayland_display EGL extension to bind
and unbind the wl_display of a wayland server to an EGLDisplay.

@feature

9 years agoecore-wayland: Check for xdg_surface before sending ack_configure
Chris Michael [Mon, 1 Dec 2014 18:52:30 +0000 (13:52 -0500)]
ecore-wayland: Check for xdg_surface before sending ack_configure

Summary: Calling wayland library functions with NULL typically leads
to an abort, so add a safety check for valid xdg_surface before
calling function

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Fix formatting
Chris Michael [Mon, 1 Dec 2014 18:51:02 +0000 (13:51 -0500)]
ecore-wayland: Fix formatting

Summary: No functional changes, just formatting

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Add safety check for xdg_popup creation
Chris Michael [Mon, 1 Dec 2014 18:47:10 +0000 (13:47 -0500)]
ecore-wayland: Add safety check for xdg_popup creation

Summary: xdg_shell_get_xdg_popup could return NULL, in which case any
calls below would cause an abort in the wayland libraries, so this
adds a trap for that.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Add safety check for window surface creation
Chris Michael [Mon, 1 Dec 2014 18:45:23 +0000 (13:45 -0500)]
ecore-wayland: Add safety check for window surface creation

Summary: We cannot call wl_proxy_get_id if the surface creation failed
else that would lead to an abort in the wayland libraries. This commit
adds a safety check

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: No need for setting the input here as it's done again
Chris Michael [Mon, 1 Dec 2014 18:43:34 +0000 (13:43 -0500)]
ecore-wayland: No need for setting the input here as it's done again
directly below.

Summary: Not much of a functional change, just removing an extra
unnecessary line. Input is being set directly below this with an if
(!( line, so no point in this one.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Check return of input_region create and get our safely
Chris Michael [Mon, 1 Dec 2014 18:15:15 +0000 (13:15 -0500)]
ecore-wayland: Check return of input_region create and get our safely

Summary: If wl_region_create fails, then we cannot call wl_region
functions on a NULL region. That causes the wayland libraries to
abort, so let's add a safety check so that we don't crash.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: If we fail to create a new opaque region, safely exit
Chris Michael [Mon, 1 Dec 2014 18:13:34 +0000 (13:13 -0500)]
ecore-wayland: If we fail to create a new opaque region, safely exit
the function.

Summary: If the compositor fails to create a new opaque region, then
we cannot call wl_region functions on a NULL region. This leads to an
abort in the wayland libraries, so let's trap the return of
wl_region_create and exit safely

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-gl: Fix evas-gl to compile for EGL upstream/1.12.0+179+gc1eb5de
Chris Michael [Mon, 1 Dec 2014 15:31:47 +0000 (10:31 -0500)]
evas-gl: Fix evas-gl to compile for EGL

Summary: Evas compilation was broken for --with-opengl=es due to the
use of GL_R16 (which is not defined for EGL).

NB: This may Not be the Proper fix, but at least it compiles now.

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoevas-gl: Fix improper #ifdef check
Chris Michael [Mon, 1 Dec 2014 15:29:53 +0000 (10:29 -0500)]
evas-gl: Fix improper #ifdef check

Summary: We should be using #ifdef here, not #if

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-wayland: Check if system has mouse in ecore_wl_input
vivek [Mon, 1 Dec 2014 15:00:19 +0000 (10:00 -0500)]
ecore-wayland: Check if system has mouse in ecore_wl_input

Summary:
Added a condition to check if system has mouse before setting
the cursor on surface.

@fix

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

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1707

9 years agoecore-wayland: Fixed a request of start_drag with null value
MinJeong Kim [Mon, 1 Dec 2014 14:58:07 +0000 (09:58 -0500)]
ecore-wayland: Fixed a request of start_drag with null value

Summary:
Fixed a request of start_drag with null value.
The request with null value has caused SIGABRT and printed
"error marshalling arguments for start_drag (signature ?oo?ou):
null value passed for arg 1".

@fix

Test Plan:
1. Run enlightenment(wayland)
2. Run elementary_test -to FeatureDnd on enlightenment
3. Click any image and try to drag it.
(Without this revision, SIGABRT will occur
 and with this revision, will not.)

Reviewers: gwanglim, devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1720

9 years agoEvas textblock: Fix possibly null dereference.
Tom Hacohen [Mon, 1 Dec 2014 11:55:08 +0000 (11:55 +0000)]
Evas textblock: Fix possibly null dereference.

Thanks to kimcinoo for reporting.

@fix

9 years agoevas: Evas_3D - add mesh blending mode.
Dmytro Dadyka [Mon, 1 Dec 2014 06:39:36 +0000 (07:39 +0100)]
evas: Evas_3D - add mesh blending mode.

Reviewers: Hermet, raster, cedric

Subscribers: Oleksander, cedric

@feature

Differential Revision: https://phab.enlightenment.org/D1548

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: Evas_3D - example which show frustum culing feature
Oleksandr Shcherbina [Thu, 9 Oct 2014 08:41:23 +0000 (11:41 +0300)]
evas: Evas_3D - example which show frustum culing feature

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: Evas_3D - add bounding sphere, revision frustum culling
Oleksandr Shcherbina [Mon, 1 Dec 2014 05:26:28 +0000 (06:26 +0100)]
evas: Evas_3D - add bounding sphere, revision frustum culling

Summary:
Move check visibility of node from evas_3d_node to evas_3d_camera
Move functionality (normalize, check distance, calculate frustum)
in evas_3d_utils.h (we are planing use evas_is_sphere_in_frustum in evas_gl_3d.c -
don't render mesh if it non visible)
Add possibility check frustum by box, aabb, central point
Refactor example frustum culling

@feature

Reviewers: Hermet, raster, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1420

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoevas: Evas_3D - fix depth texture size.
Dmytro Dadyka [Mon, 1 Dec 2014 04:57:00 +0000 (05:57 +0100)]
evas: Evas_3D - fix depth texture size.

Reviewers: raster, Hermet, cedric

Reviewed By: cedric

Subscribers: cedric

The texture used to store the depth map should be a single-channel texture.

@fix

Differential Revision: https://phab.enlightenment.org/D1713

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
9 years agoCheck return value of udev_device_get_sysattr_value(); before using atoi();
Shinwoo Kim [Sat, 29 Nov 2014 06:53:24 +0000 (15:53 +0900)]
Check return value of udev_device_get_sysattr_value(); before using atoi();

Summary: udev_device_get_sysattr_value(); could return NULL, and using atoi(); with NULL make system crash

Test Plan: It is hard to make a case to retun NULL. I got this from aging test of a product.

Reviewers: cedric, raster, seoz, woohyun, Hermet, jaehwan

Subscribers: cedric, seoz

Differential Revision: https://phab.enlightenment.org/D1687

9 years ago Use intrinsics for scaling up instead of inline asm
Carsten Haitzler (Rasterman) [Sat, 29 Nov 2014 06:50:03 +0000 (15:50 +0900)]
 Use intrinsics for scaling up instead of inline asm

Summary: Rewrite linline assembly in scaling func using NEON intrinsics.

Reviewers: raster

Differential Revision: https://phab.enlightenment.org/D1666

9 years agoevas - gl - fix warning for unused var in glx build
Carsten Haitzler (Rasterman) [Sat, 29 Nov 2014 03:07:33 +0000 (12:07 +0900)]
evas - gl - fix warning for unused var in glx build

9 years agoecore_evas_x: Fixed strange condition.
Hosang Kim [Fri, 28 Nov 2014 16:10:39 +0000 (01:10 +0900)]
ecore_evas_x: Fixed strange condition.

Summary:
To satisfy this condition, ee->visible will be EINA_FALSE. But when iconify is requested, ee->visible is usually EINA_TRUE.

@fix

Reviewers: raster, Hermet, seoz

Reviewed By: seoz

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1710

9 years agoelua: update eolian bindings
Daniel Kolesa [Fri, 28 Nov 2014 12:57:52 +0000 (12:57 +0000)]
elua: update eolian bindings

9 years agoeolian: new APIs for filenames
Daniel Kolesa [Fri, 28 Nov 2014 12:54:11 +0000 (12:54 +0000)]
eolian: new APIs for filenames

Adds 4 new APIs for getting .eo and .eot file names and file paths.

9 years agoevas/map: remove old comments.
ChunEon Park [Fri, 28 Nov 2014 11:33:49 +0000 (20:33 +0900)]
evas/map: remove old comments.

It's been so long. even SLP is not a valid name anymore.
No idea whether the problem still exist or not.
If it is then it should be reported and fixed.

9 years agoeet - add new api to verify eet file against stored cert
Carsten Haitzler (Rasterman) [Fri, 28 Nov 2014 08:54:39 +0000 (17:54 +0900)]
eet - add new  api to verify eet file against stored cert

this api makes it far more obvious as to how to verify an eet file via
the eet identify mechanisms that use x509 certificates to sign files.
this is consistent with the api used to generate the key for sigining
thus you can use the same certificate file to compare against for
identify.

@feature

9 years agoEvas GL: Use strncat and strncpy for safety
Jean-Philippe Andre [Fri, 28 Nov 2014 02:21:43 +0000 (11:21 +0900)]
Evas GL: Use strncat and strncpy for safety

Fixes Coverity reports:
- CID 1256183

Coverity was a bit stupid there. It knows the size of both
strings and complained about unsafe strcpy. It should have
complained about unsafe strcat instead.

9 years agoevas: fix build warning.
ChunEon Park [Fri, 28 Nov 2014 04:29:00 +0000 (13:29 +0900)]
evas: fix build warning.

here const pointer access is no problem.

9 years agoevas_map: Remove unnecessary check for current map
Jaehyun Cho [Fri, 28 Nov 2014 04:24:24 +0000 (13:24 +0900)]
evas_map: Remove unnecessary check for current map

Summary:
Remove unnecessary check for current map
         @fix

Reviewers: Hermet

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1708

9 years agoevas_object_main: Keep map effect after evas object move
Jaehyun Cho [Fri, 28 Nov 2014 04:18:16 +0000 (13:18 +0900)]
evas_object_main: Keep map effect after evas object move

Summary:
Keep map effect after evas object move
@feature

Reviewers: raster, cedric, Hermet

Subscribers: raster, cedric

Differential Revision: https://phab.enlightenment.org/D1678

9 years agoeolian: remove support for function types
Daniel Kolesa [Thu, 27 Nov 2014 17:20:21 +0000 (17:20 +0000)]
eolian: remove support for function types

These won't be needed because of Eo callbacks. They're also difficult
to handle in bindings, so this will relieve bindings of some effort.

9 years agoecore_con: move Ecore_Con_Dns_Cb back to C
Daniel Kolesa [Thu, 27 Nov 2014 17:13:20 +0000 (17:13 +0000)]
ecore_con: move Ecore_Con_Dns_Cb back to C

We'll be removing function pointer support from Eolian, instead
replacing any callback we can with events (arbitrary callbacks are
very difficult to support in bindings). As we'll be handling all
callbacks at once, we'll do this one at that point as well.

9 years agoEvas: Fix shader gen script for Ubuntu
Jean-Philippe Andre [Thu, 27 Nov 2014 12:46:14 +0000 (21:46 +0900)]
Evas: Fix shader gen script for Ubuntu

dash was the shell on that machine.
Also, GCC complained about a non const initializer.

Thanks SeoZ for the report.

9 years agoEvas GL: Fix up some more desktop GL extensions
Jean-Philippe Andre [Thu, 27 Nov 2014 11:46:10 +0000 (20:46 +0900)]
Evas GL: Fix up some more desktop GL extensions

OpenGL 1.2 already supports some of the features that
GLESv2 has as extensions:
- GL_EXT_read_format_bgra
- GL_EXT_texture_format_BGRA8888
- GL_EXT_texture_type_2_10_10_10_REV

Also, we need to check the proper ARB name of some extensions when
running on desktop, instead of their OES/IMG/EXT equivalent:
- GL_ARB_texture_float
- GL_ARB_texture_half_float
- GL_ARB_texture_non_power_of_two
- GL_ARB_half_float_vertex
- GL_EXT_packed_depth_stencil

9 years agoEvas GL: Also add read bgra for desktop GL
Jean-Philippe Andre [Thu, 27 Nov 2014 11:15:40 +0000 (20:15 +0900)]
Evas GL: Also add read bgra for desktop GL

9 years agoEvas GL: Advertise support for npot on desktop
Jean-Philippe Andre [Thu, 27 Nov 2014 11:08:10 +0000 (20:08 +0900)]
Evas GL: Advertise support for npot on desktop

The extension name is GL_ARB_texture_non_power_of_two
for desktop GL, but GL_OES_texture_npot for GLES.

We will consider the extensions compatible, I believe
the GLES version is a subset of the desktop one. Not sure
if that's 100% true.

9 years agoEvas GL: Always support BGRA8888 on desktop GL
Jean-Philippe Andre [Thu, 27 Nov 2014 10:47:41 +0000 (19:47 +0900)]
Evas GL: Always support BGRA8888 on desktop GL

9 years agoEvas GL: Also return fake GLSL version string
Jean-Philippe Andre [Thu, 27 Nov 2014 10:05:26 +0000 (19:05 +0900)]
Evas GL: Also return fake GLSL version string

9 years agoEvas GL: Always return "OpenGL ES 2.0" as version
Jean-Philippe Andre [Thu, 27 Nov 2014 08:17:03 +0000 (17:17 +0900)]
Evas GL: Always return "OpenGL ES 2.0" as version

This time it's for desktop GL.

I decided to keep some vendor info, as it can help for debugging
purposes.

9 years agoEvas GL: Print GLSL version with EVAS_GL_INFO
Jean-Philippe Andre [Thu, 27 Nov 2014 08:09:22 +0000 (17:09 +0900)]
Evas GL: Print GLSL version with EVAS_GL_INFO

9 years agoEvas GL: Return supported extensions list in glGetString()
Jean-Philippe Andre [Thu, 27 Nov 2014 08:03:04 +0000 (17:03 +0900)]
Evas GL: Return supported extensions list in glGetString()

9 years agoREADME: Update version number to latest
Stefan Schmidt [Thu, 27 Nov 2014 07:56:57 +0000 (08:56 +0100)]
README: Update version number to latest

Spotted by Lionel Orry. Thanks!

9 years agoEvas GL: Doc: Tell people not to read info from Evas_Native_Surface
Jean-Philippe Andre [Thu, 27 Nov 2014 07:15:23 +0000 (16:15 +0900)]
Evas GL: Doc: Tell people not to read info from Evas_Native_Surface

The information set there is not meant to be read by applications
and reused there. This would break the whole concept of letting
Evas deal with its surfaces.

Ideally Evas GL should not expose texture or fbo ID to external
applications. We could even use a new (non public) surface type
for that purpose.

9 years agoEvas GL: Return current framebuffer based on ctx information
Jean-Philippe Andre [Thu, 27 Nov 2014 05:51:58 +0000 (14:51 +0900)]
Evas GL: Return current framebuffer based on ctx information

ctx->current_fbo should contain the visible FBO value, even
if indirect rendering is used.

9 years agoEvas GL: Remove useless field
Jean-Philippe Andre [Thu, 27 Nov 2014 05:45:34 +0000 (14:45 +0900)]
Evas GL: Remove useless field

- fbo_attached is not used.
- buffers_allocated is just a bool

9 years agoEolian lua: Fix typo in function name
Jean-Philippe ANDRE [Sat, 22 Nov 2014 09:32:52 +0000 (18:32 +0900)]
Eolian lua: Fix typo in function name

9 years agoEvas GL: Remove unnecessary check
Jean-Philippe ANDRE [Thu, 20 Nov 2014 15:59:19 +0000 (00:59 +0900)]
Evas GL: Remove unnecessary check

We already checked that re->func.get_pixels wasn't NULL.

9 years agoecore-drm: pass appropriate arguments to eldbus_message_arguments_get(), so that...
Seunghun Lee [Wed, 26 Nov 2014 15:31:06 +0000 (10:31 -0500)]
ecore-drm: pass appropriate arguments to eldbus_message_arguments_get(), so that it can be return as success in _cb_device_resumed().

Summary: it seems eldbus_message_arguments_get() should be received all arguments to return success.

@fix

Reviewers: devilhorns, gwanglim

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1703

9 years agoecore-drm: Fixup formatting from D1696 patch
Chris Michael [Wed, 26 Nov 2014 14:08:55 +0000 (09:08 -0500)]
ecore-drm: Fixup formatting from D1696 patch

Signed-off-by: Chris Michael <cp.michael@samsung.com>
9 years agoecore-drm: Sending Active to login1.Session
kabeer khan [Wed, 26 Nov 2014 14:07:46 +0000 (09:07 -0500)]
ecore-drm: Sending Active to login1.Session

Summary:
Resolved TODO to set Active and State property of login1.Session
Set Active = true and State = 'active' via eldbus_proxy_property_set

@fix

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

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1696

9 years agoecore-drm: close fd handle if _ecore_drm_tty_setup() is failed in ecore_drm_tty_open().
Seunghun Lee [Wed, 26 Nov 2014 14:01:47 +0000 (09:01 -0500)]
ecore-drm: close fd handle if _ecore_drm_tty_setup() is failed in ecore_drm_tty_open().

Summary: added code to close fd handle and clear data.

@fix

Test Plan: N/A

Reviewers: gwanglim, devilhorns

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1701

9 years agoecore-drm: remove unnecessary setting file's flag of O_RDWR that are ignored.
Seunghun Lee [Wed, 26 Nov 2014 13:59:30 +0000 (08:59 -0500)]
ecore-drm: remove unnecessary setting file's flag of O_RDWR that are ignored.

Summary: setting flag of O_RDWR by F_SETFL is no effect. it's ignored.

@fix

Reviewers: gwanglim, devilhorns

Reviewed By: devilhorns

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D1695

9 years agoEvas GL: Fix indirect rendering surfaces for GLES 1.1
Jean-Philippe Andre [Wed, 26 Nov 2014 13:42:08 +0000 (22:42 +0900)]
Evas GL: Fix indirect rendering surfaces for GLES 1.1

Carefully select the requested EGL config and match it with
the available visual from X, including the following options:
- Stencil
- Depth
- MSAA

TODO: The same thing for GLX. And fix direct rendering as well.

9 years agoedje entry: improve selection performance
Thiep Ha [Wed, 26 Nov 2014 11:29:08 +0000 (11:29 +0000)]
edje entry: improve selection performance

Summary:
The selection performance is slow if we select large chunk of text.
This is caused by many rectangles created and deleted.
This patch provides a way to improve it by combine selection rectangles
of line in middle into one rectangles (i.e, if we have N lines,
the selection rectangle for lines 2 to N-1 will be combined into one.)

@feature

Reviewers: raster, cedric, tasn

Subscribers: herdsman, woohyun, cedric

Differential Revision: https://phab.enlightenment.org/D1508