platform/upstream/efl.git
8 years agoevas_gl_shader: refractor shader/program compile check code.
Amitesh Singh [Wed, 2 Dec 2015 07:47:27 +0000 (16:47 +0900)]
evas_gl_shader: refractor shader/program compile check code.

Summary:
    Call respective shader/program functions on shader/program target only.

Reviewers: cedric, raster, wonsik, spacegrapher, jpeg

Reviewed By: jpeg

Subscribers: alok25, sachin.dev, cedric, seoz

Differential Revision: https://phab.enlightenment.org/D3388

8 years agoedje cc fix - we parsed floats not intsfor min and max desc size
Carsten Haitzler (Rasterman) [Wed, 2 Dec 2015 06:24:21 +0000 (15:24 +0900)]
edje cc fix - we parsed floats not intsfor min and max desc size

thanks conr2d for pointing this out - we uses the float parse func and
not in parse. we should use int parse as the min and max sizes are
just ints and not floats. :)

@fix

8 years agoevas: refactor model's savers and loaders.
perepelits.m [Wed, 2 Dec 2015 00:38:48 +0000 (16:38 -0800)]
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)

Reviewers: cedric, raster, Hermet

Subscribers: artem.popov

Differential Revision: https://phab.enlightenment.org/D3373

8 years agoefl: fix build failure when it builds with coverage.
Youngbok Shin [Tue, 1 Dec 2015 23:30:04 +0000 (15:30 -0800)]
efl: fix build failure when it builds with coverage.

Summary:
When src/bin/efl/ builds with [--with-tests=coverage], it fails to find gcov lib.
The gcov/lcov related options has to be passed when it is builded.
And it only contained in EFL_CFLAGS, EFL_LIBS.
@fix

Test Plan:
Be sure the your enviroments to build src/bin/efl/.

1. Run ./autogen.sh --with-tests=coverage
2. make or make check
3. See the build errors.

Reviewers: raster, jpeg, cedric

Reviewed By: cedric

Subscribers: herdsman

Differential Revision: https://phab.enlightenment.org/D3370

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: fix a NULL dereference issue in font.
Youngbok Shin [Tue, 1 Dec 2015 23:03:27 +0000 (15:03 -0800)]
evas: fix a NULL dereference issue in font.

Summary:
eina_list_remove returns Eina_List pointer.
It could be NULL if the last list item is removed.
And the returned Eina_List pointer could be different from the given list.
So, calling free for fdir->data after fdir's address is changed is dangerous.
@fix

Test Plan: Run expedite or test app with evas_font_path_append() API.

Reviewers: stefan_schmidt, jpeg

Reviewed By: jpeg

Subscribers: stefan, jiin.moon, cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3392

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoedje: add mising closing brace and adjust indents and spacing in test edc
Jee-Yong Um [Tue, 1 Dec 2015 22:54:15 +0000 (14:54 -0800)]
edje: add mising closing brace and adjust indents and spacing in test edc

Summary: A closing brace is missing, and indent is not aligned.

Reviewers: cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3387

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevil: better check of Windows 64 bits
Vincent Torri [Thu, 29 Oct 2015 09:41:22 +0000 (10:41 +0100)]
evil: better check of Windows 64 bits

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: fix live resize of windows
Jean Guyomarc'h [Fri, 16 Oct 2015 08:19:57 +0000 (10:19 +0200)]
ecore_cocoa: fix live resize of windows

This was a tricky little bastard!
When a window is live resized, the NSWindow which is the target of
the live resize will wait for a kevent from the window manager,
until live resizing is done. So... live resizing is synchronous
and blocks the main thread... hence ecore_main_loop.

- When live resize starts, the Ecore_Timer which polls NSRunLoop
is paused.
- When the window is resized, the ecore_main_loop is run manually
with ecore_main_loop_iterate() to process Ecore events (mostly
Ecore_Evas)
- When live resize finished, the Ecore_Timer which polls NSRunLoop
is resumed.

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa_app: don't override init
Jean Guyomarc'h [Fri, 16 Oct 2015 07:53:26 +0000 (09:53 +0200)]
ecore_cocoa_app: don't override init

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: don't make ecore_cocoa_feed_events() a public API
Jean Guyomarc'h [Thu, 15 Oct 2015 12:06:13 +0000 (14:06 +0200)]
ecore_cocoa: don't make ecore_cocoa_feed_events() a public API

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: NSApp must be configured in the App wrapper
Jean Guyomarc'h [Thu, 15 Oct 2015 12:02:44 +0000 (14:02 +0200)]
ecore_cocoa: NSApp must be configured in the App wrapper

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa_window: cosmetic changes
Jean Guyomarc'h [Thu, 15 Oct 2015 11:58:52 +0000 (13:58 +0200)]
ecore_cocoa_window: cosmetic changes

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: don't export Ecore_Cocoa_Keys.h
Jean Guyomarc'h [Thu, 15 Oct 2015 09:45:41 +0000 (11:45 +0200)]
ecore_cocoa: don't export Ecore_Cocoa_Keys.h

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_x: factorize multiple calls to strlen()
Jean Guyomarc'h [Thu, 15 Oct 2015 09:29:51 +0000 (11:29 +0200)]
ecore_x: factorize multiple calls to strlen()

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: factorize code
Jean Guyomarc'h [Thu, 15 Oct 2015 09:13:53 +0000 (11:13 +0200)]
ecore_cocoa: factorize code

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas_gl_cocoa: make sure focus is always unlocked after locking it
Jean Guyomarc'h [Thu, 15 Oct 2015 08:47:42 +0000 (10:47 +0200)]
evas_gl_cocoa: make sure focus is always unlocked after locking it

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa_cnp: make API thread-safe
Jean Guyomarc'h [Thu, 15 Oct 2015 08:41:38 +0000 (10:41 +0200)]
ecore_cocoa_cnp: make API thread-safe

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa_cnp: clear the clipboard only if writable
Jean Guyomarc'h [Thu, 15 Oct 2015 08:38:53 +0000 (10:38 +0200)]
ecore_cocoa_cnp: clear the clipboard only if writable

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: NSEvent modifiers are of type NSUInteger
Jean Guyomarc'h [Wed, 14 Oct 2015 11:39:39 +0000 (13:39 +0200)]
ecore_cocoa: NSEvent modifiers are of type NSUInteger

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: factorize code
Jean Guyomarc'h [Wed, 14 Oct 2015 11:26:16 +0000 (13:26 +0200)]
ecore_evas_cocoa: factorize code

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: report error when [super init] fails
Jean Guyomarc'h [Wed, 14 Oct 2015 08:27:05 +0000 (10:27 +0200)]
ecore_cocoa: report error when [super init] fails

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: remove dead code
Jean Guyomarc'h [Wed, 14 Oct 2015 07:44:15 +0000 (09:44 +0200)]
ecore_evas_cocoa: remove dead code

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: remove EINA_UNUSED because parameter is used
Jean Guyomarc'h [Wed, 14 Oct 2015 07:41:08 +0000 (09:41 +0200)]
ecore_cocoa: remove EINA_UNUSED because parameter is used

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: remove debug log
Jean Guyomarc'h [Wed, 14 Oct 2015 07:40:07 +0000 (09:40 +0200)]
ecore_cocoa: remove debug log

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: reindent file
Jean Guyomarc'h [Wed, 14 Oct 2015 07:32:20 +0000 (09:32 +0200)]
ecore_cocoa: reindent file

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: add more safety checks
Jean Guyomarc'h [Wed, 14 Oct 2015 07:31:45 +0000 (09:31 +0200)]
ecore_cocoa: add more safety checks

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: remove EINA_UNUSED because parameter is actually used
Jean Guyomarc'h [Wed, 14 Oct 2015 07:18:06 +0000 (09:18 +0200)]
ecore_cocoa: remove EINA_UNUSED because parameter is actually used

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: prevent Ecore_Cocoa_Window to be re-defined
Jean Guyomarc'h [Wed, 14 Oct 2015 07:08:38 +0000 (09:08 +0200)]
ecore_cocoa: prevent Ecore_Cocoa_Window to be re-defined

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: remove ECORE_COCOA_EVENT_EXPOSE
Jean Guyomarc'h [Tue, 13 Oct 2015 16:25:01 +0000 (18:25 +0200)]
ecore_cocoa: remove ECORE_COCOA_EVENT_EXPOSE

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: remove unused events
Jean Guyomarc'h [Tue, 13 Oct 2015 16:21:36 +0000 (18:21 +0200)]
ecore_cocoa: remove unused events

8 years agoecore_cocoa: start doxygen
Jean Guyomarc'h [Tue, 13 Oct 2015 16:14:13 +0000 (18:14 +0200)]
ecore_cocoa: start doxygen

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: focus: handle ignore_events
Jean Guyomarc'h [Tue, 13 Oct 2015 15:12:53 +0000 (17:12 +0200)]
ecore_evas_cocoa: focus: handle ignore_events

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: don't track NSAppKitDefined event
Jean Guyomarc'h [Tue, 13 Oct 2015 15:09:28 +0000 (17:09 +0200)]
ecore_cocoa: don't track NSAppKitDefined event

Currently, it was used only to get track of focus, howver focus was
already handled (and better) in ecore_cocoa_window (with
NSWindowDelegate).
This led to the LOST_FOCUS event to be posted twice, with different
values, which messed up the focus stack.

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: don't reset the init count if negative
Jean Guyomarc'h [Tue, 13 Oct 2015 14:10:42 +0000 (16:10 +0200)]
ecore_evas_cocoa: don't reset the init count if negative

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: use Eina macro to determine the size of a static array
Jean Guyomarc'h [Tue, 13 Oct 2015 14:10:00 +0000 (16:10 +0200)]
ecore_evas_cocoa: use Eina macro to determine the size of a static array

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: don't shutdown ecore_event_evas() twice
Jean Guyomarc'h [Tue, 13 Oct 2015 14:08:22 +0000 (16:08 +0200)]
ecore_evas_cocoa: don't shutdown ecore_event_evas() twice

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: EAPI cannot be set for WIN32
Jean Guyomarc'h [Tue, 13 Oct 2015 14:07:38 +0000 (16:07 +0200)]
ecore_evas_cocoa: EAPI cannot be set for WIN32

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: fix types
Jean Guyomarc'h [Tue, 13 Oct 2015 13:20:27 +0000 (15:20 +0200)]
ecore_cocoa: fix types

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_evas_cocoa: remove dead code
Jean Guyomarc'h [Tue, 13 Oct 2015 12:57:52 +0000 (14:57 +0200)]
ecore_evas_cocoa: remove dead code

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: copy and paste support
Jean Guyomarc'h [Fri, 9 Oct 2015 15:18:03 +0000 (17:18 +0200)]
ecore_cocoa: copy and paste support

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: silent warning: unused parameter
Jean Guyomarc'h [Tue, 13 Oct 2015 09:05:04 +0000 (11:05 +0200)]
ecore_cocoa: silent warning: unused parameter

8 years agoecore_cocoa: use EAPI in implementation
Jean Guyomarc'h [Tue, 13 Oct 2015 09:01:15 +0000 (11:01 +0200)]
ecore_cocoa: use EAPI in implementation

8 years agoecore_cocoa: improve const-correctness
Jean Guyomarc'h [Tue, 13 Oct 2015 08:59:13 +0000 (10:59 +0200)]
ecore_cocoa: improve const-correctness

8 years agoecore_cocoa: don't specify inline
Jean Guyomarc'h [Mon, 12 Oct 2015 16:49:46 +0000 (18:49 +0200)]
ecore_cocoa: don't specify inline

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_cocoa: add support for system cursors
Jean Guyomarc'h [Thu, 8 Oct 2015 09:58:46 +0000 (11:58 +0200)]
ecore_cocoa: add support for system cursors

- Ecore_Cocoa_Cursor enum which references system cursors;
- API to show/hide cursor: ecore_cocoa_window_cursor_show();
- API to set system cursor: ecore_cocoa_window_cursor_set();
- Ecore_Evas interface to get Ecore_Cocoa_Window from Ecore_Evas.

@feature

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_win32: add support of X11 shaped cursors
Vincent Torri [Fri, 16 Oct 2015 05:08:09 +0000 (07:08 +0200)]
ecore_win32: add support of X11 shaped cursors

@feature

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_win32: add clipboard management
Vincent Torri [Thu, 15 Oct 2015 21:11:28 +0000 (23:11 +0200)]
ecore_win32: add clipboard management

@feature

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: use eina_file_path_join() and eina_environment_tmp_get() when needed
Vincent Torri [Thu, 15 Oct 2015 12:41:04 +0000 (14:41 +0200)]
eina: use eina_file_path_join() and eina_environment_tmp_get() when needed

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_file: fix ecore_file_file_get() on Windows
Vincent Torri [Thu, 15 Oct 2015 12:36:17 +0000 (14:36 +0200)]
ecore_file: fix ecore_file_file_get() on Windows

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: change efl_gfx_shape_cubic_to() api signature to follow other API (cairo and...
Subhransu Mohanty [Mon, 23 Nov 2015 01:43:46 +0000 (10:43 +0900)]
efl: change efl_gfx_shape_cubic_to() api signature to follow other API (cairo and freetype).

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: updated efl_gfx_shape_dup() implementation to use direct access to internal...
Subhransu Mohanty [Wed, 18 Nov 2015 06:27:59 +0000 (15:27 +0900)]
efl: updated efl_gfx_shape_dup() implementation to use direct access to internal data.

As I don't see the use of inheritance for stroke anymore, we will use internal access
rather than inherited get/set function. This behavior can be reverted if anyone find
a real use case for it.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: add bookkeeping for convex shapes.
Subhransu Mohanty [Fri, 6 Nov 2015 06:30:49 +0000 (15:30 +0900)]
efl: add bookkeeping for convex shapes.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: update append_rect implementation.
Subhransu Mohanty [Fri, 6 Nov 2015 01:47:37 +0000 (10:47 +0900)]
efl: update append_rect implementation.

This now generate more optimized path by reducing the use of arc and
switching to only line whenever possible.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: update append_circle implementation to use _efl_gfx_shape_append_arc.
Subhransu Mohanty [Thu, 5 Nov 2015 10:32:08 +0000 (19:32 +0900)]
efl: update append_circle implementation to use _efl_gfx_shape_append_arc.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: add alternative append_arc implementation
Subhransu Mohanty [Thu, 5 Nov 2015 10:29:55 +0000 (19:29 +0900)]
efl: add alternative append_arc implementation

This allow to reduce the number of generated arc, but doesn't follow
SVG specification. This is just used internally and can't be used by
outside call.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: add new api eina_bezier_on_interval()
Subhransu Mohanty [Wed, 25 Nov 2015 05:09:59 +0000 (14:09 +0900)]
eina: add new api eina_bezier_on_interval()

I couldn't come up with a better name, if anyone has a suggestion.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: fix bad use of calloc.
Cedric BAIL [Mon, 30 Nov 2015 22:10:35 +0000 (14:10 -0800)]
eina: fix bad use of calloc.

8 years agoeina: fix missing header
Jean Guyomarc'h [Fri, 16 Oct 2015 13:14:22 +0000 (15:14 +0200)]
eina: fix missing header

This caused functions from eina_util to be undefined at compile time.
The compiler would make implicit casts of return values into ints,
which had major side effects (e.g. segfault edje_cc)

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas_3d_utils: do not bitwise AND against 0 in condition
Stefan Schmidt [Mon, 30 Nov 2015 16:22:42 +0000 (17:22 +0100)]
evas_3d_utils: do not bitwise AND against 0 in condition

We need to check against the state here and if the compilers assignes 0 to the
first item in an enum we are screwed here as the bitwise AND will always
evaluate to false.

This is a re-incarnation from a486671bce76e405a8d7261f19b0f67ca678f0c4

8 years agoevas_font_dir: do not passed free'd memory to eina_list_remove()
Stefan Schmidt [Mon, 30 Nov 2015 16:16:28 +0000 (17:16 +0100)]
evas_font_dir: do not passed free'd memory to eina_list_remove()

Free the actual memory after we removed it from the list.

8 years agoEvas: use LC_MESSAGES on Windows to fix compilation
Vincent Torri [Fri, 27 Nov 2015 17:47:12 +0000 (18:47 +0100)]
Evas: use LC_MESSAGES on Windows to fix compilation

Compilation is broken on Windows because LC_MESSAGES does not exist on this platform
Include Evil.h to provide support of LC_MESSAGES

@fix

8 years agoEvil: add support for LC_MESSAGES for setlocale()
Vincent Torri [Fri, 27 Nov 2015 17:43:54 +0000 (18:43 +0100)]
Evil: add support for LC_MESSAGES for setlocale()

LC_MESSAGES is an extension to C ANSI and does not exist on Windows.
So add LC_MESSAGES API and overload setlocale() to support it

@feature

8 years agoecore-drm: Add logical pointer x, y variable in seat for reflecting multiple pointer...
Duna Oh [Mon, 30 Nov 2015 16:05:07 +0000 (11:05 -0500)]
ecore-drm: Add logical pointer x, y variable in seat for reflecting multiple pointer's movement

Summary: When one pointer moves, we should update the position of other devices.

Test Plan:
(1) Two pointer devices are connected.
(2) Move the cursor to (x, y) position using "device 1".
(3) When you move the cursor using "device 2", the cursor doesn't start from (x, y) position. This causes discontinuous mouse motion.

Reviewers: raster, zmike, gwanglim, stefan_schmidt, devilhorns, ManMower

Reviewed By: devilhorns, ManMower

Subscribers: cedric, Jeon, input.hacker, jpeg

Differential Revision: https://phab.enlightenment.org/D3384

8 years agoemotion: correct argument order for calloc()
Stefan Schmidt [Mon, 30 Nov 2015 15:23:00 +0000 (16:23 +0100)]
emotion: correct argument order for calloc()

calloc() expects count first and the actual size to allocate as second
argument. Say Thank You to smatch for finding this issues for us.

8 years agoexamples ecore_buffer: correct argument order for calloc()
Stefan Schmidt [Mon, 30 Nov 2015 15:22:02 +0000 (16:22 +0100)]
examples ecore_buffer: correct argument order for calloc()

calloc() expects count first and the actual size to allocate as second
argument. Say Thank You to smatch for finding this issues for us.

8 years agoevas-3d: Fix missing field initializers
Chris Michael [Mon, 30 Nov 2015 14:15:53 +0000 (09:15 -0500)]
evas-3d: Fix missing field initializers

Clang spits warnings here about missing field initializers for
Evas_Vec3 (missing y and z fields), so we will explicitly initialize
them to 0 (matching the x field).

@fix

Signed-off-by: Chris Michael <cp.michael@samsung.com>
8 years agoEvil: fix warning
Vincent Torri [Fri, 27 Nov 2015 17:48:55 +0000 (18:48 +0100)]
Evil: fix warning

Use the USERPROFILE environment variable instead of deprecated evil_homedir_get
function. Also set the shell to cmd.exe if the SHELL var is not found

8 years agoeolian: fix eo makefile example of --eo option.
JinsolPark [Mon, 30 Nov 2015 10:37:10 +0000 (10:37 +0000)]
eolian: fix eo makefile example of --eo option.

Summary: eolian: fix eo makefile example of --eo option.

Reviewers: cedric, q66

Reviewed By: q66

Subscribers: jpeg, q66

Differential Revision: https://phab.enlightenment.org/D3389

8 years agoRevert "Evas: Temporarily hack OT_SUPPORT to fix textgrid"
Tom Hacohen [Fri, 27 Nov 2015 12:34:15 +0000 (12:34 +0000)]
Revert "Evas: Temporarily hack OT_SUPPORT to fix textgrid"

Fixed in upstream, see T2865.

This reverts commit 8ccea8233c144f723470da09a01487484c269440.

8 years agoeina_matrix: optimize multiply and inverse calls by actually using the shortcut
Stefan Schmidt [Thu, 26 Nov 2015 23:13:04 +0000 (00:13 +0100)]
eina_matrix: optimize multiply and inverse calls by actually using the shortcut

We had this nice shortcuts for multiply and inverse with the identity matrix.
Pity we never used it! The EINA_MATRIX_TYPE_IDENTITY is coming from an enum
without and direct assignments to its internals. Being the first item in the
enum it is most likely will be 0 which makes the whole bitwise AND zero and thus
the optimized path will never get called. If our compiler now decides hew wants
to handle enums differently and does not assign the 0 to the first item this
bitwise operation will be even more screwed. What we really want is to check is
if the type we get for the matrix matches EINA_MATRIX_TYPE_IDENTITY. So better
do this. Made me look into matrix multply and inverse. Fun!

Thanks to smatch for poiting this out.

8 years agorg_etc: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:27:01 +0000 (17:27 +0100)]
rg_etc: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoevas software_x11: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:26:36 +0000 (17:26 +0100)]
evas software_x11: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoevas gl_common: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:26:07 +0000 (17:26 +0100)]
evas gl_common: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoevas_common3d: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:25:36 +0000 (17:25 +0100)]
evas_common3d: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoeolian: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:25:06 +0000 (17:25 +0100)]
eolian: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoector: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:24:37 +0000 (17:24 +0100)]
ector: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoevas_cserve2: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:23:47 +0000 (17:23 +0100)]
evas_cserve2: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoedje_decc: use void to force empty function parameters
Stefan Schmidt [Thu, 26 Nov 2015 16:20:14 +0000 (17:20 +0100)]
edje_decc: use void to force empty function parameters

We have to use void in a function declaration if we want no function
parameters. Using just empty parenthesis means the function takes an
unspecified number of parameters.

We had it correct for most declarations and this series fixes it for
the rest.

8 years agoevas: Add ExtraLight, ExtraBold for font weight.
Youngbok Shin [Thu, 26 Nov 2015 08:16:13 +0000 (10:16 +0200)]
evas: Add ExtraLight, ExtraBold for font weight.

Summary:
Evas supports UltraLight, UltraBold as font weight.
These terms have same weight value as ExtraLight, ExtraBold.
Some applications, for example, fontforge, use ExtraLight, ExtraBold terms for these weight values.
So, it would be better to support these terms, too.
@feature

Test Plan: None

Reviewers: tasn, woohyun, herdsman

Reviewed By: herdsman

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D3126

8 years agoedje entry: return correct selection
Thiep Ha [Thu, 26 Nov 2015 07:52:48 +0000 (09:52 +0200)]
edje entry: return correct selection

Summary:
In entry, when selection_get function is called, selection is not
always returned the current selection.

Scenario:
- In select mode, entry has selection (e.g, by double click).
- When selection handler is moved, set the cursor the current coordinate
  with edje_object_part_text_cursor_pos_set() API.
- Edje emits "selection,changed" signal.
- Elementary gets current selection and stores it.
- Elementary does not get selection as visual selection (e.g, text is
   highlighted with "entry test", but the selection returned
   from edje is "entry").
- If we copy and then paste to the entry, the pasted text is not same as
   selected text.

Reason:
- In _edje_entry_cursor_coord_set function, if entry has selection, we only
  emit "selection,changed" signal without freeing en->selection.
- When _edje_entry_selection_get is called, we check the en->selection,
  since it is existed, we just return it which is not updated one.

This patch clears en->selection, so that it is updated at _selection_get,
and the updated selection is returned to caller.

@fix

Test Plan:
In mobile profile, open entry
- Right click, choose select, double click -> selection handlers are shown.
- Drag selection handlers to change selection.
- Right click, do copy.
- Right click, do paste.
- See the pasted text is not same as selection.

Reviewers: raster, tasn, herdsman

Subscribers: seoz, JackDanielZ, cedric

Differential Revision: https://phab.enlightenment.org/D2746

8 years agobuild: rename Eolian helper cmake file to mark it as input
Stefan Schmidt [Wed, 25 Nov 2015 15:46:04 +0000 (16:46 +0100)]
build: rename Eolian helper cmake file to mark it as input

Align it with the rest of our cmake support files. Adding it to configure as
well so the cmake file gets generate from the cmake.in

Without this I had distcheck failing with a missing target for it.

8 years agocmake: Include module per default
Marcel Hollerbach [Wed, 25 Nov 2015 15:10:27 +0000 (16:10 +0100)]
cmake: Include module per default

if someone uses eolian its likly that he wants to install some rules

8 years agocmake: Add a helper script which defines generator rules for eo files
Marcel Hollerbach [Wed, 25 Nov 2015 11:22:59 +0000 (12:22 +0100)]
cmake: Add a helper script which defines generator rules for eo files

eo_rule_create adds rules to generate to .eo.x .eo.h .eot.h files.
It also monitors its deps.

@feature

8 years agoefreet_icon: Fix wrong inherit selection
Marcel Hollerbach [Sat, 21 Nov 2015 14:17:47 +0000 (15:17 +0100)]
efreet_icon: Fix wrong inherit selection

Summary:
The list of elem->paths is given with the actual icon at position 0 and
the bigger the index is the lower the icon in the inherit structure is.

Due to the for loop beginning at 0 walking to the end this
direction is flipped. So the last r is returned, which is the lowest
icon in the inherit structure.

This is fixed by returning if the first valid path is found.

@fix

Test Plan: run jesus or efm with a custom icon theme beore the wrong icons are taken, now the correct ones are taken

Reviewers: raster, cedric

Subscribers: DaveMDS

Differential Revision: https://phab.enlightenment.org/D3366

8 years agoEvas: Temporarily hack OT_SUPPORT to fix textgrid
Jean-Philippe Andre [Wed, 25 Nov 2015 06:42:29 +0000 (15:42 +0900)]
Evas: Temporarily hack OT_SUPPORT to fix textgrid

See T2865.
Since Harfbuzz 1.1.0, terminology displays fonts funnily aligned to
the top. This is apparently because until 1.0.6 the y_offset was
always 0 for all glyphs, but since 1.1.1 the offset is actually
set.

This is a TEMPORARY fix. There might be an underlying issue left
here.

Harfbuzz changed behaviour in this commit:

  commit 44f82750807475aa5b16099ccccd917d488df703
  Author: Behdad Esfahbod <behdad@behdad.org>
  Date:   Wed Nov 4 20:40:05 2015 -0800

      [ft] Remove font funcs that do nothing

8 years agoector: cleanup freetype raster
Subhransu Mohanty [Thu, 29 Oct 2015 08:20:09 +0000 (17:20 +0900)]
ector: cleanup freetype raster

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoector: support fill rule for shape object.
Subhransu Mohanty [Wed, 18 Nov 2015 05:45:16 +0000 (14:45 +0900)]
ector: support fill rule for shape object.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: added fill_rule property in efl_gfx_shape class
Subhransu Mohanty [Wed, 18 Nov 2015 05:25:09 +0000 (14:25 +0900)]
efl: added fill_rule property in efl_gfx_shape class

Open question, should fill rule be part of the public data ? Let's
say no for now.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: make path stroke related api as non virtual.
Subhransu Mohanty [Wed, 18 Nov 2015 02:13:00 +0000 (11:13 +0900)]
efl: make path stroke related api as non virtual.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoefl: quick code cleanup.
Cedric BAIL [Tue, 24 Nov 2015 22:57:58 +0000 (14:57 -0800)]
efl: quick code cleanup.

8 years agoector: updated the freetype raster with faster line renderer.
Subhransu Mohanty [Thu, 29 Oct 2015 07:59:51 +0000 (16:59 +0900)]
ector: updated the freetype raster with faster line renderer.

Merged from freetype repo with 'smooth', faster, alternative line renderer.

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoector: fix the rounding issue when applying transformation to shape data in freetype...
Subhransu Mohanty [Wed, 21 Oct 2015 07:00:28 +0000 (16:00 +0900)]
ector: fix the rounding issue when applying transformation to shape data in freetype backend.

@fix

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoevas: avoid double eet_close call while caching GL shader.
Cedric BAIL [Tue, 24 Nov 2015 22:33:20 +0000 (14:33 -0800)]
evas: avoid double eet_close call while caching GL shader.

8 years agoeet: improve error message during eet_close.
Cedric BAIL [Tue, 24 Nov 2015 22:27:13 +0000 (14:27 -0800)]
eet: improve error message during eet_close.

8 years agoEdje: Add example for hyphenation style
Daniel Hirt [Thu, 19 Nov 2015 14:25:04 +0000 (16:25 +0200)]
Edje: Add example for hyphenation style

A small example how hyphenation is set as a style in TEXTBLOCK parts.

8 years agoEvas textblock: add hyphenation tests
Daniel Hirt [Thu, 12 Nov 2015 07:49:08 +0000 (09:49 +0200)]
Evas textblock: add hyphenation tests

8 years agoEvas textblock: add support for hyphenation wrap style
Daniel Hirt [Tue, 13 Oct 2015 07:59:19 +0000 (10:59 +0300)]
Evas textblock: add support for hyphenation wrap style

We now support hyphenation in style. Use "wrap=hyphenation" to use this
wrap option. It will hyphenate based on explicit SOFT HYPHEN (&shy;)
placement in the text, and with the (optional) assistance of dictionaries
compatible with Hunspell's "hyphen" library.

This wrap mode favors breaking at hyphen positions in a word, over moving
the whole word to the next line. It will put an additional "-" at the
break position if it was hyphened.

Enabling the hyphen dictionaries is done by adding these configure
options:
  --enable-hyphen (requires Hunspell's "hyphen" library installed)
  --with-dictionaries-hyphen-dir=DIR (specifies the install location of
          the actual .dic dictionary files e.g. /usr/share/hyphen)

Note that dictionary files are expected to be in the form of "en_US.dic"
or anything that ends with it e.g. "hyph_en_US.dic" (this how they are
        named in Arch Linux).

@feature

8 years agoEina unicode: add eina_unicode_unicode_to_utf8_range
Daniel Hirt [Sat, 21 Nov 2015 10:39:00 +0000 (12:39 +0200)]
Eina unicode: add eina_unicode_unicode_to_utf8_range

Required some special treatment to get words (substring) out of an
input Unicode string to a utf8 one. This saves the trouble converting
the whole string: you input an offset in the Eina_Unicode array, and
provide the required length. That's is, now you can extract words and
whatnot in utf8 form.

To save code I wrapped the original one to get the whole length.

8 years agoeina: Applied NULL check
Anand [Tue, 24 Nov 2015 08:50:09 +0000 (17:50 +0900)]
eina: Applied NULL check

Summary:
In _eina_list_setup_accounting function
{
...
...
if (!list->accounting)
     goto on_error;
...
...
on_error:
   _eina_list_mempool_list_free(list);
}

_eina_list_mempool_list_free function deference the "list->accounting"
variable which is already NULL.

Reviewers: JackDanielZ, jpeg

Reviewed By: jpeg

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3376

8 years agoEvas 3d: Initialized local variables
Deepjyoti Dutta [Tue, 24 Nov 2015 08:03:58 +0000 (17:03 +0900)]
Evas 3d: Initialized local variables

Summary:
The OpenGL functions  assigns a value to the local variable. However,
in case of error it would take the uninitialized garbage value. So, the
correct solution would be to initialize it to 0.

Signed-off-by: Deepjyoti Dutta <deepjyoti.d@samsung.com>
Reviewers: Hermet, singh.amitesh, raster, jpeg

Reviewed By: jpeg

Subscribers: yashu21985, mvsovani, alok25, sachin.dev, cedric

Differential Revision: https://phab.enlightenment.org/D3375