platform/upstream/efl.git
8 years agoEvas filters: Fix inconsistency in transform API
Jean-Philippe Andre [Fri, 22 Jan 2016 04:54:21 +0000 (13:54 +0900)]
Evas filters: Fix inconsistency in transform API

This is a minor change, makes 'output' the default target for
the transform filter. This is consistent with all the other
filters. This is not exactly a bug fix but it doesn't break
compatibility with the earlier explicit form and improves the API
for 1.17.

8 years agoEvas filters: Disable debug mode!
Jean-Philippe Andre [Fri, 22 Jan 2016 05:38:32 +0000 (14:38 +0900)]
Evas filters: Disable debug mode!

Ooops, I must have let this in by mistake.

8 years agoEolian Cxx: Fix warnings related to deprecated EO_EV_DEL
Jean-Philippe Andre [Fri, 22 Jan 2016 06:27:35 +0000 (15:27 +0900)]
Eolian Cxx: Fix warnings related to deprecated EO_EV_DEL

Switch the generator to EO_BASE_EVENT_DEL.

8 years agoector: fix potential crash if cairo is not found on the system.
Cedric BAIL [Thu, 21 Jan 2016 23:56:49 +0000 (15:56 -0800)]
ector: fix potential crash if cairo is not found on the system.

@fix

T2938

8 years agoecore-drm: Unify and cleanup tty code
Chris Michael [Thu, 21 Jan 2016 16:38:07 +0000 (11:38 -0500)]
ecore-drm: Unify and cleanup tty code

This patch brings the tty code more inline with what the logind code
Was doing. It also fixes the return value from the vt_signal handler,
and opens the tty in non-blocking mode now.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-drm: Don't unset logind variable when disconnecting a device
Chris Michael [Thu, 21 Jan 2016 16:35:45 +0000 (11:35 -0500)]
ecore-drm: Don't unset logind variable when disconnecting a device

For a case where we have more than one device open, if we set the
'logind' variable to false, then the second device would never close
properly.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-drm: Destroy any existing eeze watch when we free the device
Chris Michael [Thu, 21 Jan 2016 16:32:53 +0000 (11:32 -0500)]
ecore-drm: Destroy any existing eeze watch when we free the device

If a user frees a device without closing it we could end up with a
dangling eeze watch. This small patch just makes sure to delete the
watch if it exists

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-drm: Refactor logind code to use tty functions
Chris Michael [Thu, 21 Jan 2016 16:20:29 +0000 (11:20 -0500)]
ecore-drm: Refactor logind code to use tty functions

This patch cleans up some of the logind code to make use of the
existing tty functions as they both did the same thing

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-evas-drm: Cleanup drm outputs during shutdown
Chris Michael [Thu, 21 Jan 2016 14:24:06 +0000 (09:24 -0500)]
ecore-evas-drm: Cleanup drm outputs during shutdown

As we no longer cleanup outputs inside the ecore_drm_device_free
function (see previous commit), we should be cleaning up outputs here
as they are created during the _ecore_evas_drm_init function.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-drm: Cleanup ecore-drm shutdown routine
Chris Michael [Thu, 21 Jan 2016 14:22:37 +0000 (09:22 -0500)]
ecore-drm: Cleanup ecore-drm shutdown routine

This patch addresses an issue where when closing Enlightenment, the
shutdown procedure would previously end up calling the same functions
twice. These functions should be called from the ecore_evas drm
shutdown routine as the sprites, inputs, outputs, etc are all called
from the ecore_evas drm init routine.

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoEo: Actually mark EO_EV_ as EINA_DEPRECATED
Jean-Philippe Andre [Thu, 21 Jan 2016 04:10:48 +0000 (13:10 +0900)]
Eo: Actually mark EO_EV_ as EINA_DEPRECATED

This should add compilation warnings whereever they're used.

8 years agoEo: Replace deprecated EO_EV_ by EO_BASE_EVENT_
Jean-Philippe Andre [Thu, 21 Jan 2016 04:10:05 +0000 (13:10 +0900)]
Eo: Replace deprecated EO_EV_ by EO_BASE_EVENT_

8 years agoEvas GL: Actually save the checksum inside the file
Jean-Philippe Andre [Thu, 21 Jan 2016 01:59:22 +0000 (10:59 +0900)]
Evas GL: Actually save the checksum inside the file

This would be cleaner than adding it as a suffix to the filename.
Use simple write and read_direct for simplicity.

This is an improvement on 3dcd903a537e29342121c and should make sure
everyone is actually using the latest version of the shaders even
if they track master or the alpha/beta releases.

8 years ago[efreet] document NULL return possibility.
Andy Williams [Wed, 20 Jan 2016 23:37:15 +0000 (23:37 +0000)]
[efreet] document NULL return possibility.

If file is not found then NULL is the fallback

8 years agobuild: fix some last corner cases for the hyphen dir being set correctly
Stefan Schmidt [Wed, 20 Jan 2016 13:11:42 +0000 (14:11 +0100)]
build: fix some last corner cases for the hyphen dir being set correctly

Making sure here that we really define EVAS_DICTS_HYPHEN_DIR after all possible
changes have been made to it and we have the final result.

8 years agoRevert "edje_cc: fix detecting missing part for insert_before/after"
Carsten Haitzler (Rasterman) [Wed, 20 Jan 2016 12:23:22 +0000 (21:23 +0900)]
Revert "edje_cc: fix detecting missing part for insert_before/after"

This reverts commit 2ca99b8b68bd52ce187c2da54944144f12afb0b6.

This breaks elm compile:

  EDJ      default.edj
edje_cc: Critical. In group "genlist_arrow": Unable to find part "sel_bevel" for insert_after in part "arrow_img_l".

8 years agoEvas: Avoid saving the shaders cache at every shutdown
Jean-Philippe Andre [Wed, 20 Jan 2016 11:33:55 +0000 (20:33 +0900)]
Evas: Avoid saving the shaders cache at every shutdown

Check the right flag.

8 years agoEvas: Discard shaders cache when the code changed
Jean-Philippe Andre [Wed, 20 Jan 2016 11:24:20 +0000 (20:24 +0900)]
Evas: Discard shaders cache when the code changed

Since we now have only two shader strings, this is trivial to do:
hash the strings and add them to the cache filename.

This will allow people using 1.17.0-alpha, etc... to discard their
old shaders cache automagically and use the latest version of the
shaders (because alpha is not in the filename).

If we end up adding more runtime generated shaders, we might need
a better strategy, but this should be good enough for now.

8 years agoEctor: Fix warning on Windows (use PRIx64 to print uint64_t)
Jean-Philippe Andre [Wed, 20 Jan 2016 10:54:44 +0000 (19:54 +0900)]
Ector: Fix warning on Windows (use PRIx64 to print uint64_t)

Thanks @vtorri for pointing this out.

8 years agobuild: sane default setting for hyphen dir
Stefan Schmidt [Wed, 20 Jan 2016 10:16:58 +0000 (11:16 +0100)]
build: sane default setting for hyphen dir

Without this compilation would fail with undefinded EVAS_DICTS_HYPHEN_DIR if
--with-xattr-tests-path was not given but hyphen enabled. Use a sane default
(/usr/share/hyphen/) to avoid this. Customization is still possible.

8 years agobuild: add hyphen as evas feature to be displayed in summary
Stefan Schmidt [Wed, 20 Jan 2016 09:44:05 +0000 (10:44 +0100)]
build: add hyphen as evas feature to be displayed in summary

Having it as a feature for evas here makes sure we can see if it is enabled
or disabled on the configure summary output.

8 years agoEvas: Actually fix this 'afill' thing
Jean-Philippe Andre [Wed, 20 Jan 2016 08:28:49 +0000 (17:28 +0900)]
Evas: Actually fix this 'afill' thing

This mostly reverts 448720fed4da252fa5eefa02ae37c79c5fb63499

After my previous patch, semi-transparent windows would render incorrectly
in E. The AFILL flag should make sure the source texture is "opaque" (ie.
if it's not marked as having an alpha channel), and masking or color
multiply should then be applied later on.

Sorry for the mess...

8 years agoEvas: Fix shader flag 'afill' (forcing alpha to 255)
Jean-Philippe Andre [Wed, 20 Jan 2016 06:57:46 +0000 (15:57 +0900)]
Evas: Fix shader flag 'afill' (forcing alpha to 255)

Apparently the shaders refactor broke the afill select logic, and it
no longer applied to objects where it should have (eg. opaque windows
in E).

8 years agoedje_cc: fix detecting missing part for insert_before/after
Jee-Yong Um [Wed, 20 Jan 2016 02:42:53 +0000 (11:42 +0900)]
edje_cc: fix detecting missing part for insert_before/after

Summary:
Detecting missing part for insert_before/after is broken.
This patch makes the feature work and clarifies error message,
and removes redundant internal function.

Reviewers: cedric, jpeg

Subscribers: jpeg

Maniphest Tasks: T2513

Differential Revision: https://phab.enlightenment.org/D3576

8 years agoedje/edje_edit: make posible to unset the program filter part and state
Vyacheslav Reutskiy [Tue, 19 Jan 2016 13:44:34 +0000 (15:44 +0200)]
edje/edje_edit: make posible to unset the program filter part and state

@fix

8 years agoEcore: Remove all temporary files after running make check
Jean-Philippe Andre [Tue, 19 Jan 2016 09:15:20 +0000 (18:15 +0900)]
Ecore: Remove all temporary files after running make check

8 years agoEina: Add proper test case for eina_mk[ds]temp
Jean-Philippe Andre [Tue, 19 Jan 2016 08:12:19 +0000 (17:12 +0900)]
Eina: Add proper test case for eina_mk[ds]temp

8 years agoEvas GL: Also fix surface cap cache
Jean-Philippe Andre [Tue, 19 Jan 2016 07:26:37 +0000 (16:26 +0900)]
Evas GL: Also fix surface cap cache

This follows the fix in eina_file_mkstemp. Not sure how it could
even work before.

@fix

8 years agoEina: Convert eina_test_file.c to UTF-8
Jean-Philippe Andre [Tue, 19 Jan 2016 06:19:51 +0000 (15:19 +0900)]
Eina: Convert eina_test_file.c to UTF-8

It was encoded as ISO-8859 (Latin-1) before.
Maybe this will break these checks on Windows, but all of EFL APIs take
UTF-8 strings, so it doesn't make sense to have ISO-8859 here.

8 years agoelua: Fix use of eina_file_mkstemp after previous patch
Jean-Philippe Andre [Tue, 19 Jan 2016 05:14:44 +0000 (14:14 +0900)]
elua: Fix use of eina_file_mkstemp after previous patch

The API (that didn't work in the first place) was used wrongly
as it was assumed its behaviour was the same as mkstemp (duh!).
It turns out eina's version doesn't replace the input string but
returns a tmpstr instead.

@fix

8 years agoEvas GL: Fix shader binaries cache
Jean-Philippe Andre [Tue, 19 Jan 2016 03:15:10 +0000 (12:15 +0900)]
Evas GL: Fix shader binaries cache

It was completetly broken due to a misuse of mkstemp.

8 years agoEina: Fix eina_file_mk[ds]temp when a path is passed
Jean-Philippe Andre [Tue, 19 Jan 2016 02:50:54 +0000 (11:50 +0900)]
Eina: Fix eina_file_mk[ds]temp when a path is passed

If the template is a path, mkstemp and mkdtemp would fail
miserably as they would try to create a file inside
/run/user/1000//path/to/file.XXXXXX even if the path did not
exist.

This patch fixes that by creating temp files inside the sys temp
dir iif the templatename is just a basic name without path
separator.

@fix

8 years agoecore-x: add safety checks for _ecore_x_disp in some functions
Mike Blumenkrantz [Mon, 18 Jan 2016 22:17:34 +0000 (17:17 -0500)]
ecore-x: add safety checks for _ecore_x_disp in some functions

xlib immediately crashes upon being passed a null DISPLAY object,
so every function in ecore-x should likely have safety checks such as
these. an easy project for anyone who has some free time

@fix

8 years agorelease: Update NEWS and bump version for 1.17.0-beta2 release
Stefan Schmidt [Mon, 18 Jan 2016 11:41:09 +0000 (12:41 +0100)]
release: Update NEWS and bump version for 1.17.0-beta2 release

8 years agoeina_cxx: make sure simple.eo lands in the tarball in all cases
Stefan Schmidt [Mon, 18 Jan 2016 12:47:39 +0000 (13:47 +0100)]
eina_cxx: make sure simple.eo lands in the tarball in all cases

This would fail if you have not enabled tests and run distcheck.

8 years agoFix build if openjpeg-2 is installed
Andrey Gursky [Mon, 18 Jan 2016 00:38:18 +0000 (01:38 +0100)]
Fix build if openjpeg-2 is installed

libopenjp2.so on Debian belongs actually to openjpeg-2.1, while
libopenjpeg.so belongs to openjpeg-1.5.

The function opj_set_default_decoder_parameters exists in both versions.
Thus another unique to openjpeg-1.5 function must be chosen to check,
whether libopenjp2 is a right one.

Fixes https://phab.enlightenment.org/T3064

8 years agoecore imf: fix imf module load hangs when not even configured
Carsten Haitzler (Rasterman) [Mon, 18 Jan 2016 08:21:50 +0000 (17:21 +0900)]
ecore imf: fix imf module load hangs when not even configured

so on some distributions scim tries to connect and hangs waiting for
connect to work. this is a problem for just loading any imf module
unless EXPLICITLY configured, so disable using an imf module unless it
is specifically configured to work. this fixes this hang. this is a
workaround though, but what else are we to do?

@fix

8 years agoembryo: change the log catagory of printf in embryo
Jaehwan Kim [Mon, 18 Jan 2016 07:54:52 +0000 (16:54 +0900)]
embryo: change the log catagory of printf in embryo

If the catagory is DBG, we should rebuild efl to print the message
by using printf in edc. So change it to use it more convenient.

@fix

8 years agowin32: Fix compilation due to recent changes in Evil
Vincent Torri [Mon, 18 Jan 2016 04:48:20 +0000 (13:48 +0900)]
win32: Fix compilation due to recent changes in Evil

Summary:
dlfcn.h is not available anymore on Windows, Evil provides all the
necessary declarations.

Reviewers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3577

@fix

8 years agoEo: Spell "Class" correctly when printing error.
Stephen Houston [Sat, 16 Jan 2016 01:41:59 +0000 (19:41 -0600)]
Eo: Spell "Class" correctly when printing error.

Leave variables named Klass so it behaves with syntax highlighting and doesn't confuse programmers.  However when Erroring, the message should be spelled correctly with "Class".

8 years agoecore: check eina_main_loop before lock.
Sung-Taek Hong [Thu, 14 Jan 2016 22:23:03 +0000 (14:23 -0800)]
ecore: check eina_main_loop before lock.

Summary:
 - EINA_MAIN_LOOP_CHECK_RETURN should be called before ecore lock
   because this may return without ecore_unlock.
 - remove EINA_UNLIKELY(!eina_main_loop_is()) which is redundant.

Reviewers: jpeg, jaehwan, cedric, raster

Reviewed By: raster

Subscribers: raster, conr2d, cedric, jpeg

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3541

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore_drm/ecore_wl2: logging: remove errant usage of %m
Derek Foreman [Thu, 14 Jan 2016 18:05:55 +0000 (13:05 -0500)]
ecore_drm/ecore_wl2: logging: remove errant usage of %m

Summary:
printf %m stringifies and prints errno.  I've tried to remove its use
anywhere that the immediately preceding function might not set errno
or is a complicated function for which knowing errno doesn't really
give any useful information.

I've left a few of the drmMode calls because they're just wrappers
around ioctl, which legitimately sets errno.

@fix

Reviewers: zmike, devilhorns

Reviewed By: zmike, devilhorns

Subscribers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3572

8 years agoecore_wayland: move check for pointer capability
Duna Oh [Thu, 14 Jan 2016 16:18:24 +0000 (11:18 -0500)]
ecore_wayland: move check for pointer capability

Summary:
We don't need to work for cursor if system doesn't have pointer devices.
Since this function is called by elementary(e.g. entry) also, we move a check condition into the function.

@fix

Reviewers: raster, zmike, devilhorns, ManMower, gwanglim

Subscribers: cedric, JHyun, input.hacker, jpeg

Differential Revision: https://phab.enlightenment.org/D3566

8 years agoeina_test: Add test-case for eina_quaternion_transform
Oleksandr Shcherbina [Thu, 14 Jan 2016 10:51:45 +0000 (11:51 +0100)]
eina_test: Add test-case for eina_quaternion_transform

Reviewers: stefan_schmidt, cedric

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3568

8 years agoecore_evas/tests: add tests for Ecore_Evas_Cocoa
Jean Guyomarc'h [Wed, 13 Jan 2016 23:09:49 +0000 (00:09 +0100)]
ecore_evas/tests: add tests for Ecore_Evas_Cocoa

8 years agoecore_evas/tests: remove trailing whitespaces
Jean Guyomarc'h [Wed, 13 Jan 2016 21:33:17 +0000 (22:33 +0100)]
ecore_evas/tests: remove trailing whitespaces

Cosmetic changes only

8 years agoevas: follow file rename to keep distcheck working
Stefan Schmidt [Thu, 14 Jan 2016 10:32:54 +0000 (11:32 +0100)]
evas: follow file rename to keep distcheck working

In 8b00eb1dd1241290025e041973f6c8748b3e8fe4 the file was moved but the rename
forgotten in the Makefile.

8 years agoEvas 3d example: Print message for usage
Jean-Philippe Andre [Thu, 14 Jan 2016 03:18:53 +0000 (12:18 +0900)]
Evas 3d example: Print message for usage

8 years agoEvas: fix example for convex hull
perepelits.m [Thu, 14 Jan 2016 02:45:26 +0000 (11:45 +0900)]
Evas: fix example for convex hull

Summary: Made new models with no normal component to print
сorrect print message for testing not only by visual result.

Reviewers: cedric, raster, Hermet

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3562

8 years agoExamples: Fix compilation of elocation ex (out of tree)
Jean-Philippe Andre [Thu, 14 Jan 2016 02:38:27 +0000 (11:38 +0900)]
Examples: Fix compilation of elocation ex (out of tree)

8 years agoemile: Added example program for base64 encoding and decoding.
Srivardhan Hebbar [Thu, 14 Jan 2016 02:28:40 +0000 (11:28 +0900)]
emile: Added example program for base64 encoding and decoding.

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

Reviewers: cedric, jpeg

Differential Revision: https://phab.enlightenment.org/D3554

8 years agoecore_wl2: fix key down event called to often
Marcel Hollerbach [Wed, 13 Jan 2016 18:24:51 +0000 (19:24 +0100)]
ecore_wl2: fix key down event called to often

if many events are coming in, a release event can be skipped. if this happens
the timer continues with the new keycode. so if a new key pressed event
enter checks for the old timer and delete the timer if there is one.

If the timer kicks in and emits additional key down events dont call the
callback which gets called by wl, and emit the signal directly.

8 years agoecore_drm: close all devices in shutdown
Marcel Hollerbach [Mon, 11 Jan 2016 21:53:18 +0000 (22:53 +0100)]
ecore_drm: close all devices in shutdown

8 years agoecore_drm: check if device is already opened before open and free
Marcel Hollerbach [Mon, 11 Jan 2016 20:43:12 +0000 (21:43 +0100)]
ecore_drm: check if device is already opened before open and free

check if fd is -1 before opening a device, so a device cannot be opened
twice.

Also check if fd is != -1 before closing it.

8 years agoeina: add all test cases for eina_vector2.
se.osadchy [Wed, 13 Jan 2016 14:43:23 +0000 (15:43 +0100)]
eina: add all test cases for eina_vector2.

Summary: All tests passes for eina_vector2 functions.

Reviewers: cedric, stefan_schmidt

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3565

8 years agoecore_drm: make sure we include math.h and link against libm for floor()
Stefan Schmidt [Wed, 13 Jan 2016 16:03:36 +0000 (17:03 +0100)]
ecore_drm: make sure we include math.h and link against libm for floor()

When we introduced the usage of floor() in ecore_drm_evdev.c we forgot to
include math.h for it as well as linking against libm. This caused linking
failures on some systems while it still worked on others. Better fix it for
good.

Original patch by aerodynamik.

ref T2953

8 years agobuild: check for dlopen() when we build the ecore_evas drm module
Stefan Schmidt [Wed, 13 Jan 2016 15:25:47 +0000 (16:25 +0100)]
build: check for dlopen() when we build the ecore_evas drm module

We use it to check for libglapi.so.0

Original patch by aerodynamik

ref T2953

8 years agoecore_wayland: send mouse_move when sets touch_focus window in cb_touch_down
Duna Oh [Wed, 13 Jan 2016 15:10:05 +0000 (10:10 -0500)]
ecore_wayland: send mouse_move when sets touch_focus window in cb_touch_down

Summary:
When user touches the screen, ecore_drm generates MOUSE_MOVE event before MOUSE_BUTTON_DOWN.
But when ecore_wayland gets touch_motion before touch_down, touch_focus window is NULL and sending MOUSE_MOVE event fails.
Also in terms of 'touch' input, touch_focus needs be set in cb_touch_down, not in cb_pointer_enter.
So this commit makes sure that ecore_wayland generates MOUSE_MOVE when touch_focus is set.

@fix

Test Plan:
when the application is launched for first time, user clicks on any area of the screen.
We can see the posion of touch event wrong.

Reviewers: raster, zmike, devilhorns, gwanglim, ManMower

Reviewed By: ManMower

Subscribers: jpeg, JHyun, cedric, input.hacker

Differential Revision: https://phab.enlightenment.org/D3555

8 years agopo: update po files
Stefan Schmidt [Wed, 13 Jan 2016 13:38:16 +0000 (14:38 +0100)]
po: update po files

8 years agoeina: Add test-case for eina_normal3_matrix_get
Oleksandr Shcherbina [Wed, 13 Jan 2016 13:14:50 +0000 (14:14 +0100)]
eina: Add test-case for eina_normal3_matrix_get

Summary: Test case fail, need apply that D3563

Reviewers: cedric, Hermet, stefan_schmidt

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3564

8 years agoeina: Fix possible wrong value due to null in determinant of matrix
Oleksandr Shcherbina [Wed, 13 Jan 2016 13:04:34 +0000 (14:04 +0100)]
eina: Fix possible wrong value due to null in determinant of matrix

Reviewers: cedric, Hermet

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3563

8 years agoeina: add test case for eina_vector2
se.osadchy [Wed, 13 Jan 2016 12:46:49 +0000 (13:46 +0100)]
eina: add test case for eina_vector2

Summary: Create eina_test_vector and add first test case.

Reviewers: cedric, stefan_schmidt

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3560

8 years agoedje: enhance edje_calc circular dependency error message
Shinwoo Kim [Wed, 13 Jan 2016 10:59:31 +0000 (19:59 +0900)]
edje: enhance edje_calc circular dependency error message

8 years agoEctor GL: Remove duplicated functions (eo)
Jean-Philippe Andre [Wed, 13 Jan 2016 08:44:01 +0000 (17:44 +0900)]
Ector GL: Remove duplicated functions (eo)

8 years agoEvas filter: Rename constructor/destructor to ctor/dtor
Jean-Philippe Andre [Wed, 13 Jan 2016 08:40:07 +0000 (17:40 +0900)]
Evas filter: Rename constructor/destructor to ctor/dtor

This avoids a name clash with Eo.Base.

8 years agoefl: edje: make comments for something we need to break for efl2 not now
Carsten Haitzler (Rasterman) [Wed, 13 Jan 2016 07:48:10 +0000 (16:48 +0900)]
efl: edje: make comments for something we need to break for efl2 not now

make notes about https://phab.enlightenment.org/D2980

8 years agoEFL: Mark EOLIAN functions as static when possible
Jean-Philippe Andre [Wed, 13 Jan 2016 06:19:07 +0000 (15:19 +0900)]
EFL: Mark EOLIAN functions as static when possible

This affects mostly evas filters.

8 years agoEvas: remove extra if from evas_3d_utils.h
perepelits.m [Wed, 13 Jan 2016 04:48:32 +0000 (13:48 +0900)]
Evas: remove extra if from evas_3d_utils.h

Summary: remove extra if from evas_3d_utils.h according to Coverity (CID1339788)

Reviewers: raster, Hermet, cedric

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3558

8 years agoEdje: enhance embryo stack broken error message
Shinwoo Kim [Wed, 13 Jan 2016 02:55:46 +0000 (11:55 +0900)]
Edje: enhance embryo stack broken error message

Summary: Enhance embry stack broken error message

Test Plan: Make a run_program run a script

Reviewers: raster, zmike, jpeg

Reviewed By: jpeg

Subscribers: cedric, seoz, jpeg

Differential Revision: https://phab.enlightenment.org/D3528

8 years agoEvas filters: Fix unused variables
Jean-Philippe Andre [Wed, 13 Jan 2016 02:11:53 +0000 (11:11 +0900)]
Evas filters: Fix unused variables

Thanks @stefan_schmidt for the report

8 years agoEvas render: Fix typo (use & instead of &)
Jean-Philippe Andre [Wed, 13 Jan 2016 02:10:25 +0000 (11:10 +0900)]
Evas render: Fix typo (use & instead of &)

Thanks @stefan_schmidt for the report

8 years agoEvas GL: Reset current surface for all contexts to null when surface is destroyed
Dongyeon Kim [Mon, 11 Jan 2016 01:11:58 +0000 (10:11 +0900)]
Evas GL: Reset current surface for all contexts to null when surface is destroyed

When evas gl surface is destroyed, not only the current surface for the current context
should be set to NULL, but current surface for all contexts should be reset.

8 years agoevas: fix resource file for billboard in evas-3d-shadow example
Oleksandr Shcherbina [Tue, 12 Jan 2016 22:05:32 +0000 (14:05 -0800)]
evas: fix resource file for billboard in evas-3d-shadow example

Summary: Due to evas.canvas3d shader flips resource image.

Reviewers: cedric, Hermet

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3556

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoedje: convert Edje_Text_Class, Edje_Size_Class data type to public
Jee-Yong Um [Tue, 12 Jan 2016 22:01:27 +0000 (14:01 -0800)]
edje: convert Edje_Text_Class, Edje_Size_Class data type to public

Summary:
APIs iterating works through active text classes and size classes were added,
but Edje_Text_Class and Edje_Size_Class data type are still concealed.
This patch uncovers those data types so as to allow developers to use them.

Reviewers: jpeg, cedric

Differential Revision: https://phab.enlightenment.org/D3557

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore-drm: Call inputs shutdown before eina_shutdown
Chris Michael [Mon, 11 Jan 2016 18:57:23 +0000 (13:57 -0500)]
ecore-drm: Call inputs shutdown before eina_shutdown

As we end up freeing the input hash inside the
_ecore_drm_inputs_shutdown function, it should be called before we
shutdown eina

@fix

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoEvas textblock: Remove redundant NULL check.
Tom Hacohen [Tue, 12 Jan 2016 10:59:52 +0000 (10:59 +0000)]
Evas textblock: Remove redundant NULL check.

Detected by smatch.

8 years agoEvas textblock: slight style fixes.
Tom Hacohen [Tue, 12 Jan 2016 10:55:49 +0000 (10:55 +0000)]
Evas textblock: slight style fixes.

8 years agoEdje UI mirroring: Fix UI mirroring for GROUP parts.
Tom Hacohen [Tue, 12 Jan 2016 10:39:37 +0000 (10:39 +0000)]
Edje UI mirroring: Fix UI mirroring for GROUP parts.

GROUP parts were not mirrored with the rest of the edje object,
this commits fixes that.

This is an improvement of the now reverted
55b62bbd70e32f5824cbc760bc2f8e49eafa4cdd.

Thanks to cedric for the pointer that lead to this commit.

Ref T3021

8 years agoRevert "Edje UI mirroring: Fix UI mirroring for GROUP parts."
Tom Hacohen [Tue, 12 Jan 2016 10:34:22 +0000 (10:34 +0000)]
Revert "Edje UI mirroring: Fix UI mirroring for GROUP parts."

The following commit solves it in a better way.

This reverts commit 55b62bbd70e32f5824cbc760bc2f8e49eafa4cdd.

8 years agoedje: unswallow objects from parent edje when swallowing into new edje
Mike Blumenkrantz [Mon, 11 Jan 2016 23:27:35 +0000 (18:27 -0500)]
edje: unswallow objects from parent edje when swallowing into new edje

the passed edje is not guaranteed to be the one which owns the swallowed
object at any given time, so get the current owner to ensure that the
correct one is passed

ref a645c8153f82c593920190416b8115b546b3668f

 #MasterOfAllBorkers

@fix

8 years agoeldbus model: no need to double const a Eina_Stringshare
Stefan Schmidt [Mon, 11 Jan 2016 21:38:25 +0000 (22:38 +0100)]
eldbus model: no need to double const a Eina_Stringshare

Eina_Stringshare is already const, no need to const it again.

8 years agoeina_inline_value: use void to force empty function parameters
Stefan Schmidt [Mon, 11 Jan 2016 21:18:37 +0000 (22:18 +0100)]
eina_inline_value: 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.

8 years agoeina-cxx: Remove Eo classes defined manually
Felipe Magno de Almeida [Mon, 11 Jan 2016 18:10:55 +0000 (16:10 -0200)]
eina-cxx: Remove Eo classes defined manually

Use Eolian to generate the Eo classes for the tests. This should fix T2940.

@fix

8 years agorelease: Update NEWS and bump version for 1.17.0-beta1 release
Stefan Schmidt [Mon, 11 Jan 2016 11:32:25 +0000 (12:32 +0100)]
release: Update NEWS and bump version for 1.17.0-beta1 release

8 years agoconfigure: Fail if opengl=full is used with EGL
Jean-Philippe Andre [Mon, 11 Jan 2016 04:47:41 +0000 (13:47 +0900)]
configure: Fail if opengl=full is used with EGL

So many problems from people who customize their builds in this
incompatible manner.

There really should be only one configure option as we don't
support opengl+egl (although possible in theory) or gles without
egl. Keeping both to not break existing builds.

8 years agoevas: fix optimization of vertex count in convex hull
perepelits.m [Fri, 8 Jan 2016 23:32:47 +0000 (00:32 +0100)]
evas: fix optimization of vertex count in convex hull

Summary: FLT_COMPARISON should be used everywhere in evas_3d_utils.h

Reviewers: cedric, raster, Hermet

Subscribers: jpeg, artem.popov

Differential Revision: https://phab.enlightenment.org/D3551

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: add test case for eina_strbuf_manage_read_only_new_length API.
Srivardhan Hebbar [Fri, 8 Jan 2016 23:28:21 +0000 (00:28 +0100)]
eina: add test case for eina_strbuf_manage_read_only_new_length API.

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

Reviewers: jpeg, cedric

Differential Revision: https://phab.enlightenment.org/D3546

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoeina: removing base64 from eina example.
Srivardhan Hebbar [Fri, 8 Jan 2016 22:51:41 +0000 (23:51 +0100)]
eina: removing base64 from eina example.

Summary:
As base64 API's are no longer part of eina, removing them. Will add
separate file for emile examples.

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

Differential Revision: https://phab.enlightenment.org/D3552

8 years agoEdje entry: Fix memory leak.
Subodh Kumar [Fri, 8 Jan 2016 16:17:29 +0000 (16:17 +0000)]
Edje entry: Fix memory leak.

Summary:
Fix memory leak

If user or extra cursor is used, these should be free on
shutdown.

@fix

Test Plan: NA

Reviewers: tasn, herdsman

Subscribers: shilpasingh, cedric, jpeg

Projects: #efl

Differential Revision: https://phab.enlightenment.org/D3550

8 years agoEdje UI mirroring: Fix UI mirroring for GROUP parts.
Tom Hacohen [Fri, 8 Jan 2016 14:57:00 +0000 (14:57 +0000)]
Edje UI mirroring: Fix UI mirroring for GROUP parts.

GROUP parts were not mirrored with the rest of the edje object,
this commits fixes that.

Fixes T3021

@fix

8 years agobuild: hopefully last fix for efl_gfx.x
Stefan Schmidt [Fri, 8 Jan 2016 10:57:55 +0000 (11:57 +0100)]
build: hopefully last fix for efl_gfx.x

Cedrics original change made sure the file was actually dist_installed but
failed to be placed inthe correct inlucde folder and thus failed the elm build.
While raster fixed this he broke the distcheck build as the file was no longer
put into the tarball being marked as nodist.
This fix handles all the cases I tried in my testing here.

8 years agoefl - fix build break with efl_gfx.x
Carsten Haitzler (Rasterman) [Fri, 8 Jan 2016 01:15:49 +0000 (10:15 +0900)]
efl - fix build break with efl_gfx.x

8 years agoedje: correct function that should have been Eo API already.
Cedric BAIL [Fri, 8 Jan 2016 00:25:17 +0000 (16:25 -0800)]
edje: correct function that should have been Eo API already.

8 years agoemile: move all eina_str_base64 to emile_base64.
Cedric BAIL [Thu, 7 Jan 2016 23:53:12 +0000 (15:53 -0800)]
emile: move all eina_str_base64 to emile_base64.

8 years agoefl: move Ector_Color to be defined in Efl.Gfx as it makes more sense.
Cedric BAIL [Thu, 7 Jan 2016 19:16:33 +0000 (11:16 -0800)]
efl: move Ector_Color to be defined in Efl.Gfx as it makes more sense.

This create some possible naming clash and is why I come up with
efl_gfx_color*_type_set functions. We will have to think about this more
carefully as it makes sense to now pass this colors directly to our color
API. Ideally the default 8 bits interface would become just a convenience
wrapper around the more complex possibility.

8 years agojs: fix binding after namespace change of connector
Lauro Moura [Thu, 7 Jan 2016 18:31:53 +0000 (10:31 -0800)]
js: fix binding after namespace change of connector

Summary:
Update register function after change introduced in
the commit dc560521509b93a58905c8280d7cf9622d4e87c4
changed the native API.

@fix

Reviewers: q66, cedric, felipealmeida

Subscribers: jpeg

Differential Revision: https://phab.enlightenment.org/D3547

Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
8 years agoecore-evas: Rename ecore_evas_wayland_window_get2 function
Chris Michael [Thu, 7 Jan 2016 20:02:13 +0000 (15:02 -0500)]
ecore-evas: Rename ecore_evas_wayland_window_get2 function

This renames the ecore_evas_wayland_window_get2 function to be
ecore_evas_wayland2_window_get before the 1.17 roll out.

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoecore-wl2: API expose ecore_wl2_input_grab function
Chris Michael [Thu, 7 Jan 2016 16:55:52 +0000 (11:55 -0500)]
ecore-wl2: API expose ecore_wl2_input_grab function

This input_grab function should have been exposed via EAPI for use in
enlightenment, but was missed before the push

Signed-off-by: Chris Michael <cpmichael@osg.samsung.com>
8 years agoeina mp: also include malloc_np.h on FreeBSD
Daniel Kolesa [Thu, 7 Jan 2016 15:43:26 +0000 (15:43 +0000)]
eina mp: also include malloc_np.h on FreeBSD

8 years agoeina mp: only include malloc.h on linux
Daniel Kolesa [Thu, 7 Jan 2016 15:31:59 +0000 (15:31 +0000)]
eina mp: only include malloc.h on linux

Elsewhere, everything is in stdlib.h.

@fix