platform/upstream/efl.git
6 years agoelementary: apply anti-aliasing at the gesture layer example.
Hermet Park [Tue, 10 Apr 2018 02:25:25 +0000 (11:25 +0900)]
elementary: apply anti-aliasing at the gesture layer example.

6 years agoevas: improve evas map anti-alising rendering quality.
Hermet Park [Tue, 10 Apr 2018 02:09:35 +0000 (11:09 +0900)]
evas: improve evas map anti-alising rendering quality.

Old version algorithm was imperfection a bit, quality was poor at some specific
degrees, specifically, when pixel increment pattern on the diagonal lines is
unstable.

This revised version was better than old one even source code is much cleaner
and simpler.

See belows.

*NonAA vs AA:
https://ibb.co/bCNfMc

*Compare the worst case aa in the old version:
https://ibb.co/bEJsZx

*Test video:
https://youtu.be/Wn20Tym5lfg

6 years agoelementary: add new config value to handle tap finger size.
Woochan Lee [Mon, 9 Apr 2018 17:27:57 +0000 (10:27 -0700)]
elementary: add new config value to handle tap finger size.

Summary:
The tap_finger_size value of the concept is different from the finger_size value used in the past.

We need a minimum value for recognize the gesture as a tap or not.

Since the actual screen has different screen sizes, there is a problem that the recognition rate of tap is lowered when using the existing value(finger_size)

Test Plan: elementary_test -> gesture sample.

Reviewers: cedric, woohyun, Jaehyun_Cho, herb, id213sin

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5842

Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
Signed-off-by: Cedric BAIL <cedric@osg.samsung.com>
6 years agoecore_evas: check return value.
Woochan Lee [Mon, 9 Apr 2018 17:18:38 +0000 (10:18 -0700)]
ecore_evas: check return value.

Summary: After migration this code in Tizen. The coverity said it needs to check return value(CID 39562).

Reviewers: raster, myoungwoon, woohyun, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5907

Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
6 years agoevas: add null check
JunsuChoi [Mon, 9 Apr 2018 17:12:58 +0000 (10:12 -0700)]
evas: add null check

Summary:
this commit add null check.
In EVAS_OBJECT_DATA_SAFE_GET, if eo_obj is null, obj can also be null.
and In _efl_canvas_output_async_block() can return null.

Test Plan: N/A

Reviewers: woohyun, kimcinoo, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5870

Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
6 years agoecore_fb: fix divided by zero
JunsuChoi [Mon, 9 Apr 2018 17:11:19 +0000 (10:11 -0700)]
ecore_fb: fix divided by zero

Summary:
w and h must not be zero.
because of the divide_by_zero exception occurs on line 649 and 653.

Test Plan: N/A

Reviewers: woohyun, kimcinoo, cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5873

Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
6 years agoefl_ui_focus_object:
Marcel Hollerbach [Fri, 6 Apr 2018 10:48:41 +0000 (12:48 +0200)]
efl_ui_focus_object:

make the prepare call safe to recursive calls. There is no need to
prepare a item again if it is already in prepare.

6 years agoefl_ui_focus_manager_root_focus: refactor _state_eval
Marcel Hollerbach [Mon, 9 Apr 2018 16:27:09 +0000 (18:27 +0200)]
efl_ui_focus_manager_root_focus: refactor _state_eval

Summary:
we have to check whether all of descendant are none_logical.

this fixes T6800

Test Plan:
1. elementary_test -to ctxpopup
2. Click 'Ctxpopup with user content'
3. Try focus ctxpopup content(button) using arrow key.
4. Check that the content is focused (you can move scroll bar)

Reviewers: bu5hm4n

Subscribers: raster, cedric

Maniphest Tasks: T6800

Differential Revision: https://phab.enlightenment.org/D5856

6 years agointroduce efl_ui_focus_manager_window_root
Marcel Hollerbach [Mon, 9 Apr 2018 14:16:16 +0000 (16:16 +0200)]
introduce efl_ui_focus_manager_window_root

Which is a interface that indicates that this is the root of all focus
managers.

6 years agoefl_ui_focus_util: support later managers not beeing registered
Marcel Hollerbach [Sat, 7 Apr 2018 21:59:47 +0000 (23:59 +0200)]
efl_ui_focus_util: support later managers not beeing registered

this should fix some corner cases when setting focus by hand

6 years agoefl_ui_focus_layer: return the registered focus manager when active
Marcel Hollerbach [Sat, 7 Apr 2018 15:25:35 +0000 (17:25 +0200)]
efl_ui_focus_layer: return the registered focus manager when active

this fixes in chain setting for hovers and ctxpopups

6 years agoelm_hoversel: make the keyevent handlers work again
Marcel Hollerbach [Sat, 7 Apr 2018 12:47:00 +0000 (14:47 +0200)]
elm_hoversel: make the keyevent handlers work again

i am not sure when this stopped working, elm hover sends focus to the
children it has, thus the propergation line for the key down events is
going from the hover up accross the over parent, however, the
hoverparent can be different to the hoversel, which would lead to
missing navigation in the hoversel.

this fixes T6802

6 years agotests: eolian: remove reference to no longer existing files
Stefan Schmidt [Mon, 9 Apr 2018 10:24:53 +0000 (12:24 +0200)]
tests: eolian: remove reference to no longer existing files

These have been renamed in commit f19a905261c6b02f788ce7b5167223555d324712
No need to keep the reference in the makefile. distcheck was also
getting confused about them.

6 years agoecore: fix a potential memory leak.
Hermet Park [Mon, 9 Apr 2018 02:46:13 +0000 (11:46 +0900)]
ecore: fix a potential memory leak.

if the func argument is null, allocated memory won't be freed properly.
return it as soon as possible.

@fix

6 years agoedje: fix potential memory leak.
Hermet Park [Mon, 9 Apr 2018 02:29:41 +0000 (11:29 +0900)]
edje: fix potential memory leak.

next_token() requires free() for it's returned value and a string attribute.
Here token usage didn't take care of the memory freeing.

@fix

6 years agoedje: fix potential memory leak.
Hermet Park [Mon, 9 Apr 2018 02:14:07 +0000 (11:14 +0900)]
edje: fix potential memory leak.

next_token() requires free() for it's returned value and a string attribute
Here logic didn't take care of the memory usage.

@fix

6 years agogendoc-mono: Add documentation generation for C# language
Felipe Magno de Almeida [Wed, 21 Feb 2018 18:40:15 +0000 (15:40 -0300)]
gendoc-mono: Add documentation generation for C# language

6 years agoefl-js: Fix dependency generation in JavaScript binding compilation
Felipe Magno de Almeida [Tue, 2 Jan 2018 14:00:31 +0000 (12:00 -0200)]
efl-js: Fix dependency generation in JavaScript binding compilation

Use C++ as dependency in automake so concurrent builds will not fail to compile

6 years agoA really small fix for test focus 5
Dave Andreoli [Sat, 7 Apr 2018 18:02:51 +0000 (20:02 +0200)]
A really small fix for test focus 5

Just place one of the buttons in the perfect position.

6 years agoRevert "efl_ui_focus_manager_root_focus: refactor _state_eval"
Carsten Haitzler (Rasterman) [Sat, 7 Apr 2018 07:34:29 +0000 (16:34 +0900)]
Revert "efl_ui_focus_manager_root_focus: refactor _state_eval"

This reverts commit c6ce04e48f7813475e929f6205a2f6a2e7eb9729.

OK. I bisected and found that this is ythe root cause for 2 pretty
chunky bugs. Reverting this fixes T6870 and fixes T6871.

so that's 2 segvs in existing uses and even olne (elm config) ships
with efl. Having a look at the diff doesn't jump out what is wrong
here, but the best course IMHO is to take this patch and put it back
into review again... :)

6 years agoevas: when passed an Evas_Object, we should try to access it directly so that during...
Cedric Bail [Fri, 6 Apr 2018 18:28:55 +0000 (11:28 -0700)]
evas: when passed an Evas_Object, we should try to access it directly so that during destructor Evas is still accessible.

6 years agoevas: fix typo written twice
JunsuChoi [Fri, 6 Apr 2018 17:16:23 +0000 (10:16 -0700)]
evas: fix typo written twice

Summary: disp is written twice with the same value.

Test Plan: N/A

Reviewers: woohyun, kimcinoo, cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5874

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelementary: in case of an object being destroyed, there is no point to warn.
Cedric Bail [Fri, 6 Apr 2018 18:31:03 +0000 (11:31 -0700)]
elementary: in case of an object being destroyed, there is no point to warn.

6 years agoecore_wayland: remove logically dead code.
Bowon Ryu [Fri, 6 Apr 2018 17:10:42 +0000 (10:10 -0700)]
ecore_wayland: remove logically dead code.

Summary:
remove logically dead code.
execution cannot reach this statement.

Reviewers: cedric

Differential Revision: https://phab.enlightenment.org/D5876

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoefl_part: remove duplicated text_aliases code
YeongJong Lee [Fri, 6 Apr 2018 17:10:05 +0000 (10:10 -0700)]
efl_part: remove duplicated text_aliases code

Summary: this widgets inherit from Layout. they can use same text_aliases of Layout.

Reviewers: Jaehyun_Cho, woohyun

Subscribers: herb, cedric

Differential Revision: https://phab.enlightenment.org/D5859

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoarc: update config to use current config value names
Mike Blumenkrantz [Fri, 6 Apr 2018 17:08:29 +0000 (10:08 -0700)]
arc: update config to use current config value names

Summary: Depends on D5867

Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5868

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoecore: reduce pipe wait time during shutdown
Mike Blumenkrantz [Fri, 6 Apr 2018 17:08:24 +0000 (10:08 -0700)]
ecore: reduce pipe wait time during shutdown

Summary:
this is only meant to listen to data which is currently available,
not wait for new data

@fix

Depends on D5866

Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5867

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoecore: poll the thread pipe twice during shutdown
Mike Blumenkrantz [Fri, 6 Apr 2018 17:08:23 +0000 (10:08 -0700)]
ecore: poll the thread pipe twice during shutdown

Summary:
this resolves a race condition when a thread join was pending during
shutdown but a pipe write was needed in order for the join to be
successfully executed before shutdown had occurred

@fix

Reviewers: cedric

Reviewed By: cedric

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5866

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoelm_inwin: use macro for content set/get/unset
YeongJong Lee [Fri, 6 Apr 2018 16:28:20 +0000 (09:28 -0700)]
elm_inwin: use macro for content set/get/unset

Summary: also, content_aliases is not necessary. it can use Layout's content_aliases.

Test Plan:
elementary_test -to inwin
elementary_test -to inwin2
elementary_test -to inwin3

Reviewers: Jaehyun_Cho, woohyun

Subscribers: herb, cedric

Differential Revision: https://phab.enlightenment.org/D5860

Reviewed-by: Cedric Bail <cedric@osg.samsung.com>
6 years agoefl_selection_manager: fix build on MacOS and WL2.
Alastair Poole [Fri, 6 Apr 2018 15:11:41 +0000 (16:11 +0100)]
efl_selection_manager: fix build on MacOS and WL2.

6 years agoefl_wl: Add NULL check
YeongJong Lee [Fri, 6 Apr 2018 14:38:10 +0000 (09:38 -0500)]
efl_wl: Add NULL check

Reviewers: zmike, ManMower, Jaehyun_Cho, woohyun

Reviewed By: ManMower

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5862

6 years agoecore audio - add comment that this is an intended leak...
Carsten Haitzler (Rasterman) [Fri, 6 Apr 2018 13:40:00 +0000 (22:40 +0900)]
ecore audio - add comment that this is an intended leak...

6 years agotests: eldbus: mark unused parameter as such
Stefan Schmidt [Fri, 6 Apr 2018 09:36:51 +0000 (11:36 +0200)]
tests: eldbus: mark unused parameter as such

6 years agotests: ecore: remove no longer needed jenkins check function
Stefan Schmidt [Fri, 6 Apr 2018 09:19:40 +0000 (11:19 +0200)]
tests: ecore: remove no longer needed jenkins check function

After changes to the ecore_timer tests this function is no longer used
or needed. It can safely go.

6 years agotests: elementary: remove unused variable
Stefan Schmidt [Thu, 5 Apr 2018 14:12:25 +0000 (16:12 +0200)]
tests: elementary: remove unused variable

Looks like a copy and paste left over from another test.

../../../src/tests/elementary/elm_test_genlist.c:58:16: warning: unused variable ‘type’ [-Wunused-variable]

6 years agoecore audio - dont unload modules etc. in case audio obj around on exit
Carsten Haitzler (Rasterman) [Fri, 6 Apr 2018 08:08:44 +0000 (17:08 +0900)]
ecore audio - dont unload modules etc. in case audio obj around on exit

if audio objects are around after ecore_audio has been shut down then
they may access the sndfile or pulse lib ptrs/funcs and thus dont
unload these. should fix crashes on shutdown.

6 years agoecore: and rename the test for Eina_Future to be the only promise test.
Cedric Bail [Wed, 4 Apr 2018 23:12:49 +0000 (16:12 -0700)]
ecore: and rename the test for Eina_Future to be the only promise test.

6 years agoecore: remove tests of old Efl_Future based on Eo.
Cedric Bail [Wed, 4 Apr 2018 22:59:41 +0000 (15:59 -0700)]
ecore: remove tests of old Efl_Future based on Eo.

6 years agoevas: by moving lifecycle management to be handle by the death of the parent, there...
Cedric Bail [Wed, 4 Apr 2018 22:36:24 +0000 (15:36 -0700)]
evas: by moving lifecycle management to be handle by the death of the parent, there should be no call to efl_del.

6 years agoevas: the life cycle of an object inside an hash can not rely on the parent destroyin...
Cedric Bail [Wed, 4 Apr 2018 21:50:42 +0000 (14:50 -0700)]
evas: the life cycle of an object inside an hash can not rely on the parent destroying, use ref/unref properly.

6 years agoecore: rely on object being destroyed on parent invalidate.
Cedric Bail [Thu, 22 Mar 2018 01:20:53 +0000 (18:20 -0700)]
ecore: rely on object being destroyed on parent invalidate.

6 years agoecore: pipe fd can become invalid while being processed.
Cedric Bail [Thu, 22 Mar 2018 01:19:24 +0000 (18:19 -0700)]
ecore: pipe fd can become invalid while being processed.

6 years agoevas: we should always sync the canvas before destroying the object.
Cedric BAIL [Thu, 5 Apr 2018 18:27:44 +0000 (11:27 -0700)]
evas: we should always sync the canvas before destroying the object.

6 years agotests: ck_assert_ptr_null -> ck_assert_ptr_eq
Mike Blumenkrantz [Thu, 5 Apr 2018 18:03:50 +0000 (14:03 -0400)]
tests: ck_assert_ptr_null -> ck_assert_ptr_eq

6 years agotests: ck_assert_ptr_null -> ck_assert_ptr_ne
Mike Blumenkrantz [Thu, 5 Apr 2018 17:43:15 +0000 (13:43 -0400)]
tests: ck_assert_ptr_null -> ck_assert_ptr_ne

6 years agoeina: mark thread by default as not cancellable as per Eina documentation.
Cedric BAIL [Thu, 5 Apr 2018 17:39:30 +0000 (10:39 -0700)]
eina: mark thread by default as not cancellable as per Eina documentation.

By default all Pthread are cancellable. This leads to strange behavior that
happen during shutdown and hasn't been figured out. Most likely fix for T6622.

6 years agotests: move bindings tests to using checked fixtures too.
Mike Blumenkrantz [Thu, 5 Apr 2018 17:15:36 +0000 (13:15 -0400)]
tests: move bindings tests to using checked fixtures too.

T6862
T6811

Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
6 years agoelementary: fix efl_ui_view_list examples
Xavi Artigas [Thu, 5 Apr 2018 16:59:54 +0000 (09:59 -0700)]
elementary: fix efl_ui_view_list examples

Reviewers: SanghyeonLee

Reviewed By: SanghyeonLee

Subscribers: cedric

Differential Revision: https://phab.enlightenment.org/D5863

Reviewed-by: Cedric BAIL <cedric@osg.samsung.com>
6 years agotests: add test for check < 0.11 to enable timing info
Mike Blumenkrantz [Thu, 5 Apr 2018 16:25:16 +0000 (12:25 -0400)]
tests: add test for check < 0.11 to enable timing info

6 years agotests: elementary: fix build break after rename of efl_ui_view examples
Stefan Schmidt [Thu, 5 Apr 2018 12:30:50 +0000 (14:30 +0200)]
tests: elementary: fix build break after rename of efl_ui_view examples

Commit ae8df9ba076e6086900353f0e10480bb2302b44d renamed the files but
did not rename the references in the Makefile.

6 years agoelm - selection manager - fix string buffer handling and format types
Carsten Haitzler (Rasterman) [Thu, 5 Apr 2018 12:12:10 +0000 (21:12 +0900)]
elm - selection manager - fix string buffer handling and format types

this has 2 bugs really. 1 - the format masks for plain text included
markup which ... is wrong - right? that should have a mime type of its
own... either way it leads to plain text pasts being handled wrong in
wayland... so this fixes that

also the string buffers in wayland just were nothandled right. buffers
were raw binary buffers plus size,not strings, but they were strdup'd
etc. assuming nul termination. oh so wrong to do that. this also fixes
that as i fpound it while fixing the above format bug.

this fixes T6284

@fix

6 years agoeolian: _ is allowed in doc references
Daniel Kolesa [Thu, 5 Apr 2018 11:53:13 +0000 (13:53 +0200)]
eolian: _ is allowed in doc references

6 years agoFix static analysis result
Shinwoo Kim [Thu, 5 Apr 2018 04:18:03 +0000 (13:18 +0900)]
Fix static analysis result

[Dereference after null check]

(1) src/lib/ecore/ecore_main.c
 - _efl_loop_handler_efl_object_finalize checks if pd->loop_data is NULL.
   After that, _handler_reset > _handler_clear > _ecore_main_fd_handler_del >
   _ecore_main_fdh_pool_del is directly dereferencing pd->pool_data.
 - _efl_loop_handler_efl_object_parent_set checks if pd->loop_data as well.
   Then it calls _handler_reset as well.

(2) src/lib/ecore_wayland/ecore_wl_dnd.c
  - ecore_wl_dnd_selection_set checks if t - result of wl_array_add - is NULL.
    And it is dereferecing t directly for wl_data_source_offer.

(3) src/lib/elementary/efl_ui_dnd.c
 - Third parameter const char *data could be NULL.
   In this case strlen dereferences NULL. The data should be non NULL value.
   I have checked this with Mr. Thiep Ha.

(4) src/lib/evas/canvas/evas_object_inform.c
 - _efl_canvas_object_efl_gfx_stack_stack_below checks if obj->layer is NULL.
   So it could call evas_object_inform_call_call_restack which is dereferencing
   obj->layer directly.

6 years agoefl selection manager - build again if no wl or no x is enabled
Carsten Haitzler (Rasterman) [Thu, 5 Apr 2018 10:16:16 +0000 (19:16 +0900)]
efl selection manager - build again if no wl or no x is enabled

so make check and my builds passed because i have wl and x enabled. if
one of these isnt then this build will fail. my fix fails. this fixes
that.

6 years agoelm - fix selection debug print to not print unterminated buffers
Carsten Haitzler (Rasterman) [Thu, 5 Apr 2018 09:50:14 +0000 (18:50 +0900)]
elm - fix selection debug print to not print unterminated buffers

so it's printfing a non-0 byte terminated string ... it's a
buffer/slice which has explicit lenmgth and no nul, so fwrite it
approplriately.

6 years agoevas - plug in dirty + pixels get hold in "is inside" for images
Carsten Haitzler (Rasterman) [Thu, 5 Apr 2018 05:38:24 +0000 (14:38 +0900)]
evas - plug in dirty + pixels get hold in "is inside" for images

this should fix T6791

@fix

6 years agoefl selection manager - handle multi window selections properly
Carsten Haitzler (Rasterman) [Wed, 4 Apr 2018 16:20:37 +0000 (01:20 +0900)]
efl selection manager - handle multi window selections properly

tell other windows they lost their selections correctly if you have
multiple windows within the same process. this fixes multi-window mode
in terminology.

@fix

6 years agoefl selection - fix paste with multiple windows and filter by window
Carsten Haitzler (Rasterman) [Wed, 4 Apr 2018 14:39:50 +0000 (23:39 +0900)]
efl selection - fix paste with multiple windows and filter by window

so changes in elm cnp (efl selection) broke cnp with terminology
multi-window mode. this fixes it.

@fix

6 years agoefl threadio - make call and call sync func proto same as efl events
Carsten Haitzler (Rasterman) [Wed, 4 Apr 2018 10:35:19 +0000 (19:35 +0900)]
efl threadio - make call and call sync func proto same as efl events

well call sync returns a void * too.... instead of just void return
(nothing) but this makes it easier to write and maintain code due to
consistency of function callback signatures.

6 years agoefl exe - map some "well known" exit codes to specific errnos
Carsten Haitzler (Rasterman) [Wed, 4 Apr 2018 09:57:55 +0000 (18:57 +0900)]
efl exe - map some "well known" exit codes to specific errnos

6 years agoefl task - support flag to ignore exit code errors for future resolve
Carsten Haitzler (Rasterman) [Tue, 3 Apr 2018 17:34:10 +0000 (02:34 +0900)]
efl task - support flag to ignore exit code errors for future resolve

6 years agoecore - thread - document thread max limits more sensibly
Carsten Haitzler (Rasterman) [Tue, 3 Apr 2018 07:41:03 +0000 (16:41 +0900)]
ecore - thread - document thread max limits more sensibly

leave the exact numbers less defined because core count can change on
the fly and could be virtually limited by policy in future etc. so
keep things less exactly defined so people dont go depending on exact
results which was never really intended.

6 years agoelm config tool - get web backend when entry is deleted not on terminate
Carsten Haitzler (Rasterman) [Mon, 2 Apr 2018 08:00:20 +0000 (17:00 +0900)]
elm config tool - get web backend when entry is deleted not on terminate

terminate may be called after widgets/windows are deleted already...
so donyt get this stuff in terminate.

6 years agoelm config - dont leak config if re-loading it again later on
Carsten Haitzler (Rasterman) [Mon, 2 Apr 2018 07:46:55 +0000 (16:46 +0900)]
elm config - dont leak config if re-loading it again later on

delete config object and so on before the relaod

6 years agoelm win - change parent to only use win parents in efl win struct as parent
Carsten Haitzler (Rasterman) [Mon, 2 Apr 2018 07:46:14 +0000 (16:46 +0900)]
elm win - change parent to only use win parents in efl win struct as parent

6 years agotests: increase step sizes for eina_fp_div and eina_fp_mul tests
Mike Blumenkrantz [Fri, 30 Mar 2018 21:08:48 +0000 (17:08 -0400)]
tests: increase step sizes for eina_fp_div and eina_fp_mul tests

fix T6841

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce looping in slstr_thread thread callbacks
Mike Blumenkrantz [Fri, 30 Mar 2018 20:58:31 +0000 (16:58 -0400)]
tests: reduce looping in slstr_thread thread callbacks

this works out to the same number of total tests as slstr_many but
now split across all the threads

fix T6846

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce looping in eina_test_spinlock
Mike Blumenkrantz [Fri, 30 Mar 2018 20:52:20 +0000 (16:52 -0400)]
tests: reduce looping in eina_test_spinlock

no changes to coverage

fix T6845

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: add eina_test_hash_int64_fuzze
Mike Blumenkrantz [Fri, 30 Mar 2018 20:14:00 +0000 (16:14 -0400)]
tests: add eina_test_hash_int64_fuzze

ref T6844

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce looping in eina_ustringshare_putstuff
Mike Blumenkrantz [Fri, 30 Mar 2018 19:45:56 +0000 (15:45 -0400)]
tests: reduce looping in eina_ustringshare_putstuff

fix T6843

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce looping in eina_binshare_putstuff test
Mike Blumenkrantz [Fri, 30 Mar 2018 19:41:33 +0000 (15:41 -0400)]
tests: reduce looping in eina_binshare_putstuff test

also slightly increase eina_hash coverage

fix T6843

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: improve eina_binshare_collision test
Mike Blumenkrantz [Fri, 30 Mar 2018 19:21:02 +0000 (15:21 -0400)]
tests: improve eina_binshare_collision test

remove rand() usage and loop less

fix T6843

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: change eina cmdline test names to match api namespaces
Mike Blumenkrantz [Fri, 30 Mar 2018 19:13:12 +0000 (15:13 -0400)]
tests: change eina cmdline test names to match api namespaces

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: remove 9900 loop iterations from eina_test_file_thread
Mike Blumenkrantz [Fri, 30 Mar 2018 19:01:04 +0000 (15:01 -0400)]
tests: remove 9900 loop iterations from eina_test_file_thread

no changes in coverage

fix T6842

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: rework eina hash fuzzing tests
Mike Blumenkrantz [Fri, 30 Mar 2018 18:27:51 +0000 (14:27 -0400)]
tests: rework eina hash fuzzing tests

these are supposed to be for testing the finding of values not in the
hash, not basic hash operations

fix T6839

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: fix mem leak in eina_test_hash_add_del_by_hash
Mike Blumenkrantz [Fri, 30 Mar 2018 17:52:46 +0000 (13:52 -0400)]
tests: fix mem leak in eina_test_hash_add_del_by_hash

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce malloc usage in eina_test_hash_extended
Mike Blumenkrantz [Fri, 30 Mar 2018 17:47:48 +0000 (13:47 -0400)]
tests: reduce malloc usage in eina_test_hash_extended

ref T6839

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reformat eina_test_hash.c
Mike Blumenkrantz [Fri, 30 Mar 2018 17:20:50 +0000 (13:20 -0400)]
tests: reformat eina_test_hash.c

no functional changes

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: use namespacing for eina hash tests
Mike Blumenkrantz [Fri, 30 Mar 2018 17:18:36 +0000 (13:18 -0400)]
tests: use namespacing for eina hash tests

no functional changes

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: use namespacing for safepointer test names
Mike Blumenkrantz [Fri, 30 Mar 2018 17:12:25 +0000 (13:12 -0400)]
tests: use namespacing for safepointer test names

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce array size in safepointer tests
Mike Blumenkrantz [Fri, 30 Mar 2018 17:10:47 +0000 (13:10 -0400)]
tests: reduce array size in safepointer tests

this reduces test runtime by ~1.0s while having no effect on coverage

fix T6836

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: simplify eina safepointer tests to be more readable
Mike Blumenkrantz [Fri, 30 Mar 2018 16:51:58 +0000 (12:51 -0400)]
tests: simplify eina safepointer tests to be more readable

no functional changes

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: use EINA_C_ARRAY_LENGTH macro in safepointer tests
Mike Blumenkrantz [Fri, 30 Mar 2018 16:34:15 +0000 (12:34 -0400)]
tests: use EINA_C_ARRAY_LENGTH macro in safepointer tests

no functional changes

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: rework strbuf append/prepend tests
Mike Blumenkrantz [Fri, 30 Mar 2018 15:11:57 +0000 (11:11 -0400)]
tests: rework strbuf append/prepend tests

there's no point in iterating this many times; the base step size is 32
and so each loop will only increase the buffer...not at all since it takes
3 loops for the buffer to increase by the step size

verify that the buffer increases and then stop instead of spinning uselessly

fix T6835

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reformat eina_test_strbuf.c
Mike Blumenkrantz [Fri, 30 Mar 2018 14:58:09 +0000 (10:58 -0400)]
tests: reformat eina_test_strbuf.c

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: use namespacing for eina strbuf tests
Mike Blumenkrantz [Fri, 30 Mar 2018 14:55:38 +0000 (10:55 -0400)]
tests: use namespacing for eina strbuf tests

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: improve eina_test_list_shuffle
Mike Blumenkrantz [Thu, 29 Mar 2018 22:14:53 +0000 (18:14 -0400)]
tests: improve eina_test_list_shuffle

reduce iteratons by a hundredfold, remove verifying of rand() seed,
check whether shuffle actually shuffles

fix T6833

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: rename eina_list tests to use list namespace for readability
Mike Blumenkrantz [Thu, 29 Mar 2018 21:39:59 +0000 (17:39 -0400)]
tests: rename eina_list tests to use list namespace for readability

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: remove sleep calls from eina_barrier_test_simple
Mike Blumenkrantz [Thu, 29 Mar 2018 21:02:04 +0000 (17:02 -0400)]
tests: remove sleep calls from eina_barrier_test_simple

the timing does not matter for these tests, so there's no point in
randomly waiting longer and delaying build completion

fix T6832

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: improve coverage in eina_rbtree_fuzzy test
Mike Blumenkrantz [Thu, 29 Mar 2018 20:51:33 +0000 (16:51 -0400)]
tests: improve coverage in eina_rbtree_fuzzy test

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: simplify eina_rbtree_fuzzy test and speedup by infinity%
Mike Blumenkrantz [Thu, 29 Mar 2018 20:43:33 +0000 (16:43 -0400)]
tests: simplify eina_rbtree_fuzzy test and speedup by infinity%

this test is to try and match values in a tree which have not been
inserted, so remove irrelevant insert/remove operations as well as
insane looping

fix T6830

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: greatly reduce iterations for eina_rbtree_remove test
Mike Blumenkrantz [Thu, 29 Mar 2018 19:54:02 +0000 (15:54 -0400)]
tests: greatly reduce iterations for eina_rbtree_remove test

this has no bearing on coverage, so there is not much point in uselessly
iterating here

ref T6830

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: reduce malloc calls in eina rbtree tests
Mike Blumenkrantz [Thu, 29 Mar 2018 19:14:34 +0000 (15:14 -0400)]
tests: reduce malloc calls in eina rbtree tests

ref T6830

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: rework eina inlist tests to be infinity% faster
Mike Blumenkrantz [Thu, 29 Mar 2018 18:41:31 +0000 (14:41 -0400)]
tests: rework eina inlist tests to be infinity% faster

coverage currently at ~95%

fix T6829

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agoelm: remove hacky calls to emotion_init/shutdown
Mike Blumenkrantz [Tue, 27 Mar 2018 20:02:39 +0000 (16:02 -0400)]
elm: remove hacky calls to emotion_init/shutdown

these are irrelevant since emotion manages init/shutdown internally

fixes elm test pass state

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agoemotion: call shutdown when smart objects are destroyed
Mike Blumenkrantz [Tue, 27 Mar 2018 20:01:53 +0000 (16:01 -0400)]
emotion: call shutdown when smart objects are destroyed

the constructor calls init for every object, so ensure a matching
shutdown call exists to balance the init count

ref 517471190f87e473e3bc38584c23ae28fb772c7c
ref T6816

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: unbreak PACKAGE_BUILD_DIR usage in makefiles
Mike Blumenkrantz [Tue, 27 Mar 2018 19:31:00 +0000 (15:31 -0400)]
tests: unbreak PACKAGE_BUILD_DIR usage in makefiles

for some reason, a few projects changed this from the expected value of
abs_top_builddir for their test suites, breaking the usage of this define
within those tests

fix T6812

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: disable xim in ecore_imf test when DISPLAY is not set
Mike Blumenkrantz [Tue, 27 Mar 2018 19:14:11 +0000 (15:14 -0400)]
tests: disable xim in ecore_imf test when DISPLAY is not set

tests should be able to run in console without failing

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agoelput: include Eina.h in public header
Mike Blumenkrantz [Tue, 27 Mar 2018 18:13:46 +0000 (14:13 -0400)]
elput: include Eina.h in public header

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>
6 years agotests: redo eina in-tree module loading
Mike Blumenkrantz [Tue, 27 Mar 2018 18:07:19 +0000 (14:07 -0400)]
tests: redo eina in-tree module loading

this performs an eina init/shutdown check and loads available modules
before running remaining tests

note that this still does not work and may never have worked because
PACKAGE_BUILD_DIR is broken

ref T6812

Reviewed-by: Stefan Schmidt <stefan@osg.samsung.com>