platform/upstream/efl.git
12 years agoAdd a new API(ecore_evas_request_geometry_get)
Jiyoun Park [Fri, 28 Oct 2011 11:15:33 +0000 (11:15 +0000)]
Add a new API(ecore_evas_request_geometry_get)

The reson why I add this is for communicate with X in async mode.
For example, If applications call elm_win_rotation_with_resize_set API
when they start run and rotation mode is set.
ecore size doesn't changed yet, so it make elm window size to 1
becaues elm window's resize function use ecore_evas_geometry_get API.
so I add new api help upperside get info related with recently request geometry

SVN revision: 64492

12 years agoFrom: "Sung W. Park" <sungwoo@gmail.com>
Sung W. Park [Fri, 28 Oct 2011 11:08:23 +0000 (11:08 +0000)]
From: "Sung W. Park" <sungwoo@gmail.com>
Subject: [E-devel] [E-Devel][Patch] Evas GL Color Format Enum change
(and ElmGLview changes accordingly)

I'm submitting a patch that changes the color format for Evas GL.

When I first wrote Evas_GL, I just had EVAS_GL_RGB_8 and EVAS_GL_RGBA_8 and etc
but it was misleading for some people.  It was more of a filler since I couldn't decide on
a name.  I'm finally changing it to make it more clear.

SVN revision: 64491

12 years agoadd the documentations about edje group inherit.
Jaehwan Kim [Fri, 28 Oct 2011 09:01:38 +0000 (09:01 +0000)]
add the documentations about edje group inherit.

SVN revision: 64488

12 years agoecore/fb: Change fb engine input to use ecore_input_evas.
Rafael Antognolli [Thu, 27 Oct 2011 19:29:08 +0000 (19:29 +0000)]
ecore/fb: Change fb engine input to use ecore_input_evas.
Nothing changes, only making the ecore fb engine to send keyboard and
mouse events using ecore_input_evas, instead of its own ecore events.

Patch for SiT.

SVN revision: 64447

12 years agoecore/ecore_fb: Disable mouse move event filtering.
Rafael Antognolli [Thu, 27 Oct 2011 18:57:23 +0000 (18:57 +0000)]
ecore/ecore_fb: Disable mouse move event filtering.
There are two main reasons for disabling this now:
 - it's not working, some problem with an assertion of ecore_lock is
 making the program abort;
 - it's wrong, mouse move events can only be filtered out if they are
 received from an absolute axis device (touchscreen and so), not
 relative ones, like mouse or touchpads.
Patch for SiT.

SVN revision: 64444

12 years agoevas - updated Changelog and @since for evas_object_freeze_events_set/get
ChunEon Park [Thu, 27 Oct 2011 13:19:52 +0000 (13:19 +0000)]
evas - updated Changelog and @since for evas_object_freeze_events_set/get

Thank you Vincent.

SVN revision: 64434

12 years agoBetter docs for ecore_main_loop_iterate()
Iván Briano [Thu, 27 Oct 2011 13:19:11 +0000 (13:19 +0000)]
Better docs for ecore_main_loop_iterate()

SVN revision: 64433

12 years agoevas - added new API evas_object_freeze_events_set/get
ChunEon Park [Thu, 27 Oct 2011 10:36:09 +0000 (10:36 +0000)]
evas - added new API evas_object_freeze_events_set/get

and will handle for the key events also soon.

SVN revision: 64432

12 years agoeio: fix maintainer-clean
Michael BOUCHAUD [Thu, 27 Oct 2011 10:17:44 +0000 (10:17 +0000)]
eio: fix maintainer-clean

SVN revision: 64431

12 years agoand fix evas map used as mapbuf with clipping and stuff! ugh! the
Carsten Haitzler [Thu, 27 Oct 2011 09:39:18 +0000 (09:39 +0000)]
and fix evas map used as mapbuf with clipping and stuff! ugh! the
forever shifting bug! evas_render needs a rewrite

SVN revision: 64429

12 years agoFrom: Jihoon Kim [mailto:jihoon48.kim@samsung.com]
Carsten Haitzler [Thu, 27 Oct 2011 05:42:41 +0000 (05:42 +0000)]
From: Jihoon Kim [mailto:jihoon48.kim@samsung.com]

elm_entry_autocapital_type_set API is high level API to call ecore_imf_context_autocapital_type_set.
Application programmer can choose the type of autocapitalization such as
WORD, SENTENCE, ALLCHARACTER through this API.

elm_entry_input_panel_enabled_set API is for setting whether input panel
(virtual keyboard) should be appeared when entry has a focus or pressed.
It can be used by dialer or calculator application programmer because those
application prefer to use its keypad NOT virtual keyboard supported by system.
They want to use entry to show the cursor for providing the cursor handling
method but don't want to appear system keyboard.

SVN revision: 64423

12 years agoEdje: minor - Fix swallow part error msg
Bruno Dilly [Wed, 26 Oct 2011 19:49:09 +0000 (19:49 +0000)]
Edje: minor - Fix swallow part error msg

SVN revision: 64416

12 years agoEvas font: Removed double ;.
Tom Hacohen [Wed, 26 Oct 2011 14:33:51 +0000 (14:33 +0000)]
Evas font: Removed double ;.

SVN revision: 64414

12 years agofix clip-out issue after making mapped objs opque fix was in.
Carsten Haitzler [Wed, 26 Oct 2011 13:33:25 +0000 (13:33 +0000)]
fix clip-out issue after making mapped objs opque fix was in.

SVN revision: 64413

12 years agowe can't limit sizes on load because of "i'm just loading the header
Carsten Haitzler [Wed, 26 Oct 2011 08:33:41 +0000 (08:33 +0000)]
we can't limit sizes on load because of "i'm just loading the header
to find out the size" tricks. :(

SVN revision: 64410

12 years agothanks to jiyoun's work in finding why this happens - here's actually a
Carsten Haitzler [Wed, 26 Oct 2011 05:29:38 +0000 (05:29 +0000)]
thanks to jiyoun's work in finding why this happens - here's actually a
simpler fix. :)

SVN revision: 64405

12 years agowarning--
Carsten Haitzler [Tue, 25 Oct 2011 12:25:15 +0000 (12:25 +0000)]
warning--

SVN revision: 64390

12 years agowarn--
Carsten Haitzler [Tue, 25 Oct 2011 12:20:53 +0000 (12:20 +0000)]
warn--

SVN revision: 64389

12 years agowarn--
Carsten Haitzler [Tue, 25 Oct 2011 12:14:48 +0000 (12:14 +0000)]
warn--

SVN revision: 64388

12 years agowarning--
Carsten Haitzler [Tue, 25 Oct 2011 12:07:56 +0000 (12:07 +0000)]
warning--

SVN revision: 64387

12 years agodisable getting of xft.dpi - always been an option but it just leads
Carsten Haitzler [Tue, 25 Oct 2011 12:01:44 +0000 (12:01 +0000)]
disable getting of xft.dpi - always been an option but it just leads
to inconsistent rendering.

SVN revision: 64386

12 years agooh dear.. lewft and right pixle duplication for interpolation had the
Carsten Haitzler [Tue, 25 Oct 2011 09:27:54 +0000 (09:27 +0000)]
oh dear.. lewft and right pixle duplication for interpolation had the
WRONG PIXELS!!! amazing it never crashed! fix fix! (also handle
opengls that doesnt support unpack row length)

SVN revision: 64377

12 years agoevas/evas_object_smart - trivial changes
ChunEon Park [Tue, 25 Oct 2011 08:37:43 +0000 (08:37 +0000)]
evas/evas_object_smart - trivial changes

SVN revision: 64375

12 years agoFrom: EunMi Lee <eunmi15.lee@samsung.com>
EunMi Lee [Tue, 25 Oct 2011 05:03:50 +0000 (05:03 +0000)]
From: EunMi Lee <eunmi15.lee@samsung.com>
Subject: [E-devel] [Patch] [Evas] Patch to provide information of
touched points

Hello,
I made a new patch to get information of current touched point instead
of Touch Event.

I added touch_points (Eina_List) to the Evas structure and it maintains touched points on the evas.
New touched point is added to the touch_points when we get Mouse_Down and Multi_Down,
touched point is updated when we get Mouse_Move and Mult_Move,
and touched point is removed when we get Mouse_Up and Multi_Up.

The each touch point has coordinate, id and state information as follows:
id - identifier. 0 for Mouse Event and device id for Multi Event. coordinate - (x, y) coordinate of point.
state - state of point. type is Evas_Touch_Point_State enum.
(EVAS_TOUCH_POINT_DOWN, EVAS_TOUCH_POINT_UP, EVAS_TOUCH_POINT_MOVE,
EVAS_TOUCH_POINT_STILL, EVAS_TOUCH_POINT_CANCEL)

There are 4 new APIs to get touch point's information as follows:
unsigned int evas_touch_point_list_count(Evas *e);
void evas_touch_point_list_nth_xy_get(Evas *e, unsigned int n, Evas_Coord *x, Evas_Coord *y);
int evas_touch_point_list_nth_id_get(Evas *e, unsigned int n);
Evas_Touch_Point_State evas_touch_point_list_nth_state_get(Evas *e, unsigned int n);

I added APIs to get each information instead of exposing whole
structure to make it easy to expand in the future as you mentioned in
the below e-mail :)

SVN revision: 64373

12 years agoEvas - Modified field names in Evas structure to be more clear.
ChunEon Park [Tue, 25 Oct 2011 02:26:51 +0000 (02:26 +0000)]
Evas - Modified field names in Evas structure to be more clear.

SVN revision: 64371

12 years agoEcore_X(cb): Remove some dead code. Fix formatting (how did this get
Christopher Michael [Sun, 23 Oct 2011 18:33:32 +0000 (18:33 +0000)]
Ecore_X(cb): Remove some dead code. Fix formatting (how did this get
sooo messed up ??).

SVN revision: 64328

12 years agoEcore imf: Free the module list on shutdown.
Tom Hacohen [Sun, 23 Oct 2011 12:01:36 +0000 (12:01 +0000)]
Ecore imf: Free the module list on shutdown.

Freeing the content (using eina_module_list_free) is not enough, we need
to also free the array itself.

SVN revision: 64326

12 years agoemotion: correctly guard Ecore_X usage.
Cedric BAIL [Sun, 23 Oct 2011 09:21:14 +0000 (09:21 +0000)]
emotion: correctly guard Ecore_X usage.

SVN revision: 64322

12 years agoemotion: fix borderless and remove it from taskbar/pager at the same time.
Cedric BAIL [Sun, 23 Oct 2011 08:46:23 +0000 (08:46 +0000)]
emotion: fix borderless and remove it from taskbar/pager at the same time.

SVN revision: 64321

12 years agoemacs users don't like trailing white spaces
Vincent Torri [Sun, 23 Oct 2011 06:59:40 +0000 (06:59 +0000)]
emacs users don't like trailing white spaces

SVN revision: 64319

12 years agoannoy vim users with spaces at the start of blank lines. :)
Carsten Haitzler [Sun, 23 Oct 2011 03:51:41 +0000 (03:51 +0000)]
annoy vim users with spaces at the start of blank lines. :)

SVN revision: 64317

12 years agoevas: Fix a warning
Mike McCormack [Sat, 22 Oct 2011 23:33:23 +0000 (23:33 +0000)]
evas: Fix a warning

  CC     evas_cserve_tool.o
evas_cserve_main.c: In function ‘message’:
evas_cserve_main.c:993:14: warning: braces around scalar initializer
evas_cserve_main.c:993:14: warning: (near initialization for ‘lopt.degree’)
evas_cserve_main.c:993:14: warning: excess elements in scalar initializer
evas_cserve_main.c:993:14: warning: (near initialization for ‘lopt.degree’)
evas_cserve_main.c:993:14: warning: excess elements in scalar initializer
evas_cserve_main.c:993:14: warning: (near initialization for ‘lopt.degree’)
evas_cserve_main.c:993:14: warning: excess elements in scalar initializer
evas_cserve_main.c:993:14: warning: (near initialization for ‘lopt.degree’)

Signed-off-by: Mike McCormack <mikem@ring3k.org>
SVN revision: 64310

12 years agoUpdated Czech translations in ecore, thanks belisarivs.
Daniel Kolesa [Sat, 22 Oct 2011 23:28:48 +0000 (23:28 +0000)]
Updated Czech translations in ecore, thanks belisarivs.

SVN revision: 64306

12 years agoEvas: bump libpng version to 1.2.10.
Vincent Torri [Sat, 22 Oct 2011 19:39:56 +0000 (19:39 +0000)]
Evas: bump libpng version to 1.2.10.
libpng 1.2.8 does not have the symbol png_set_expand_gray_1_2_4_to_8.
It seems that 1.2.10 has no problem, so we check for libpng >= 1.2.10
and we drop libpng 1.0.*

SVN revision: 64303

12 years agoEcore_X(cb): Make sure we have argv[i] before we try to strlen it.
Christopher Michael [Sat, 22 Oct 2011 18:06:44 +0000 (18:06 +0000)]
Ecore_X(cb): Make sure we have argv[i] before we try to strlen it.

SVN revision: 64302

12 years agoand fix chglog and @since for max img size api
Carsten Haitzler [Sat, 22 Oct 2011 05:42:48 +0000 (05:42 +0000)]
and fix chglog and @since for max img size api

SVN revision: 64290

12 years agoand zero out previous rect info on clear.
Carsten Haitzler [Sat, 22 Oct 2011 03:56:36 +0000 (03:56 +0000)]
and zero out previous rect info on clear.

SVN revision: 64289

12 years agoand fix tiler to do full resolution. yes - more mem, some more
Carsten Haitzler [Sat, 22 Oct 2011 03:36:39 +0000 (03:36 +0000)]
and fix tiler to do full resolution. yes - more mem, some more
overhead, but fewer bugs.

SVN revision: 64288

12 years agooh dear. debugging snuck in in my other fixes. remove.
Carsten Haitzler [Sat, 22 Oct 2011 02:40:52 +0000 (02:40 +0000)]
oh dear. debugging snuck in in my other fixes. remove.

SVN revision: 64287

12 years agooh this one was a doosey to find. objects still having map applied BUT
Carsten Haitzler [Fri, 21 Oct 2011 17:14:03 +0000 (17:14 +0000)]
oh this one was a doosey to find. objects still having map applied BUT
map is the same as a normal display (no mapping at all) are not
considered opaque. fix!

SVN revision: 64264

12 years agoevas/fb: use virtual sizes to describe the fb mode geometry.
Rafael Antognolli [Fri, 21 Oct 2011 16:21:09 +0000 (16:21 +0000)]
evas/fb: use virtual sizes to describe the fb mode geometry.
This was causing the fb engine to draw incorrectly when the virtual size
was different from the size (xres != xres_virtual). For example, it
happens an additional monitor is plugged into a notebook.
Patch for SiT.

SVN revision: 64259

12 years agoecore_evas_fb: use _ecore_evas_fb_render instead of the buffer one.
Rafael Antognolli [Fri, 21 Oct 2011 16:07:29 +0000 (16:07 +0000)]
ecore_evas_fb: use _ecore_evas_fb_render instead of the buffer one.
This mistake was introduced long ago on a cleanup.
Patch for SiT.

SVN revision: 64258

12 years agoemotion: disable direct rendering path if window manager doesn't handle it.
Cedric BAIL [Fri, 21 Oct 2011 15:00:41 +0000 (15:00 +0000)]
emotion: disable direct rendering path if window manager doesn't handle it.

SVN revision: 64256

12 years agoemotion: try to correctly fix the fill value.
Cedric BAIL [Fri, 21 Oct 2011 14:59:22 +0000 (14:59 +0000)]
emotion: try to correctly fix the fill value.

SVN revision: 64255

12 years agoevas: coorectly dirty YUV data.
Cedric BAIL [Fri, 21 Oct 2011 14:06:42 +0000 (14:06 +0000)]
evas: coorectly dirty YUV data.

SVN revision: 64254

12 years agohmm gles doesnt have internal format fetching.
Carsten Haitzler [Fri, 21 Oct 2011 12:03:41 +0000 (12:03 +0000)]
hmm gles doesnt have internal format fetching.

SVN revision: 64253

12 years agoand NOW fix obscure object handling properly.
Carsten Haitzler [Fri, 21 Oct 2011 10:25:35 +0000 (10:25 +0000)]
and NOW fix obscure object handling properly.

SVN revision: 64252

12 years agoevas/evas_inline - indent!
ChunEon Park [Fri, 21 Oct 2011 10:10:47 +0000 (10:10 +0000)]
evas/evas_inline - indent!

SVN revision: 64251

12 years agonote on detecting texture alloc fail. need to percolate that back
Carsten Haitzler [Fri, 21 Oct 2011 10:03:17 +0000 (10:03 +0000)]
note on detecting texture alloc fail. need to percolate that back

SVN revision: 64250

12 years agoand disable invalidate optimization
Carsten Haitzler [Fri, 21 Oct 2011 10:00:03 +0000 (10:00 +0000)]
and disable invalidate optimization

SVN revision: 64249

12 years agodebug for future. for now found the problem.
Carsten Haitzler [Fri, 21 Oct 2011 09:59:13 +0000 (09:59 +0000)]
debug for future. for now found the problem.

SVN revision: 64248

12 years agoevas/evas_events - trivial changes
ChunEon Park [Fri, 21 Oct 2011 09:53:26 +0000 (09:53 +0000)]
evas/evas_events - trivial changes

SVN revision: 64247

12 years agohave evas engine also to update rect math so we can exclude useless
Carsten Haitzler [Fri, 21 Oct 2011 08:58:00 +0000 (08:58 +0000)]
have evas engine also to update rect math so we can exclude useless
updates when obscured objects change

SVN revision: 64246

12 years agoadd call to get maximum image size (eg max texture size)
Carsten Haitzler [Fri, 21 Oct 2011 08:17:14 +0000 (08:17 +0000)]
add call to get maximum image size (eg max texture size)

SVN revision: 64244

12 years agoif image biger than max texture size... report allocation errors.
Carsten Haitzler [Fri, 21 Oct 2011 06:10:15 +0000 (06:10 +0000)]
if image biger than max texture size... report allocation errors.

SVN revision: 64241

12 years agonicer formatting of async events
Carsten Haitzler [Fri, 21 Oct 2011 05:50:05 +0000 (05:50 +0000)]
nicer formatting of async events

SVN revision: 64240

12 years agoecrustify on ecore-x.
Carsten Haitzler [Fri, 21 Oct 2011 05:40:59 +0000 (05:40 +0000)]
ecrustify on ecore-x.

SVN revision: 64239

12 years agoran ecrustify on ecore. some manual fixing too.
Carsten Haitzler [Fri, 21 Oct 2011 05:40:39 +0000 (05:40 +0000)]
ran ecrustify on ecore. some manual fixing too.

SVN revision: 64238

12 years agoran ecrustify on eet. this seems to do a good job of eet. comments?
Carsten Haitzler [Fri, 21 Oct 2011 05:40:01 +0000 (05:40 +0000)]
ran ecrustify on eet. this seems to do a good job of eet. comments?

SVN revision: 64237

12 years agoevas/evas_events.c - removed duplicated lines
ChunEon Park [Thu, 20 Oct 2011 12:21:41 +0000 (12:21 +0000)]
evas/evas_events.c - removed duplicated lines

SVN revision: 64198

12 years agoevas/evas_events - removed useless lines since we support multi downs
ChunEon Park [Thu, 20 Oct 2011 11:02:22 +0000 (11:02 +0000)]
evas/evas_events - removed useless lines since we support multi downs

but looks bad to keep the only one point in evas.

SVN revision: 64197

12 years agolets have envrionment var control over semisync in ecore-evas
Carsten Haitzler [Thu, 20 Oct 2011 06:39:03 +0000 (06:39 +0000)]
lets have envrionment var control over semisync in ecore-evas

SVN revision: 64191

12 years agoEcore: do not enter po/ if --disable-nls is passed to configure
Vincent Torri [Wed, 19 Oct 2011 17:21:33 +0000 (17:21 +0000)]
Ecore: do not enter po/ if --disable-nls is passed to configure

SVN revision: 64183

12 years agoeina: fallback for case where file system doesn't handle MAP_HUGETLB properly.
Cedric BAIL [Wed, 19 Oct 2011 16:09:14 +0000 (16:09 +0000)]
eina: fallback for case where file system doesn't handle MAP_HUGETLB properly.

SVN revision: 64181

12 years agowe dont need the dirty at the end either.
Carsten Haitzler [Wed, 19 Oct 2011 09:50:55 +0000 (09:50 +0000)]
we dont need the dirty at the end either.

SVN revision: 64173

12 years agoemotion: destroy all created object.
Cedric BAIL [Wed, 19 Oct 2011 09:42:05 +0000 (09:42 +0000)]
emotion: destroy all created object.

SVN revision: 64172

12 years agoFrom: Jaehwan Kim <jae.hwan.kim@samsung.com>
Jaehwan Kim [Wed, 19 Oct 2011 09:41:17 +0000 (09:41 +0000)]
From: Jaehwan Kim <jae.hwan.kim@samsung.com>
Subject: [E-devel] [Patch] change rel1.to to default when using group
inherit

When using group inherit, if rel1.to is set in parent group, we cannot change the part name
of rel1.to to default (the default is that dont have the part name) in child group.
So I changed edje_cc_out.c. After patching, we can change rel1.to to
default by using "". (ex > rel1.to: "";)

SVN revision: 64171

12 years agoFrom: Jiyoun Park <jy0703.park@samsung.com>
Jiyoun Park [Wed, 19 Oct 2011 09:04:34 +0000 (09:04 +0000)]
From: Jiyoun Park <jy0703.park@samsung.com>
Subject: [E-devel] [Patch] Add scale down decoding feature to evas png loader

I add scale down decoding feature to evas png loader. 5515X3986 size png image need 80~90M memory,
but scale down(scale num=2) option can reduce memory to 25~30M.

I use down sample method for scale down.
(there is more efficient algorithm for scale down, I'll add this to my
todo list)

SVN revision: 64170

12 years agoevas: simplify code at the same time.
Cedric BAIL [Wed, 19 Oct 2011 09:04:18 +0000 (09:04 +0000)]
evas: simplify code at the same time.

SVN revision: 64169

12 years agonono - we dont dirty images when we set data...
Carsten Haitzler [Wed, 19 Oct 2011 08:59:29 +0000 (08:59 +0000)]
nono - we dont dirty images when we set data...

SVN revision: 64168

12 years agoeina/stringshare - get rid of white space
ChunEon Park [Wed, 19 Oct 2011 08:36:42 +0000 (08:36 +0000)]
eina/stringshare - get rid of white space

SVN revision: 64165

12 years agoFrom: Jiyoun Park <jy0703.park@samsung.com>
Jiyoun Park [Wed, 19 Oct 2011 06:36:28 +0000 (06:36 +0000)]
From: Jiyoun Park <jy0703.park@samsung.com>
Subject: Re: [E-devel] [Patch] Implement scale down decoding feature of bmp loader

I implement scale down decoding feature of bmp loader using down sample algorithm.
Desktop have low risk to go wrong memory problem during big image decoding,
but mobile device is different.

Raster said it is life (meet memory problem during big image decoding),
and it is enough to return decoding fail.

But I think it seems a bit harsh, because 2 or 3 bmp image (5000X5000 size: for example)
can cause application close because of memory lack.

SVN revision: 64163

12 years agoEvas font: Oops, hb_fonts should be per font instance, not source.
Tom Hacohen [Tue, 18 Oct 2011 09:58:49 +0000 (09:58 +0000)]
Evas font: Oops, hb_fonts should be per font instance, not source.

SVN revision: 64144

12 years agoFrom: Jiyoun Park <jy0703.park@samsung.com>
Jiyoun Park [Tue, 18 Oct 2011 09:10:26 +0000 (09:10 +0000)]
From: Jiyoun Park <jy0703.park@samsung.com>
Subject: [E-devel] [Patch] modify gl engine's animated function
related with cache entry

I modified the gl engine code related with animated images
This is very trivial.  Evas image object passes images to the engine.
In the software engine, it is a cache entry , but in the GL engine, it is
an Evas_GL Image. So I modified the gl engine code to get the cache entry
from the gl image.

SVN revision: 64143

12 years agofor now - revers eumni touch events patch until we get press state
Carsten Haitzler [Tue, 18 Oct 2011 08:32:03 +0000 (08:32 +0000)]
for now - revers eumni touch events patch until we get press state
"getter" api patch. :)

SVN revision: 64140

12 years agoFrom: "Sung W. Park" <sungwoo@gmail.com>
Sung W. Park [Tue, 18 Oct 2011 08:13:23 +0000 (08:13 +0000)]
From: "Sung W. Park" <sungwoo@gmail.com>
Subject: Re: [E-devel] [E-Devel][Review][Patch] Evas GL Extensions + a
bug fix

Here's an initial attempt at the GL extensions issue for Evas GL.

I have been in discussion with a few EFL developers regarding how we should
provide extensions.  Essentially, there are two ways to go about doing this.

1. provide evas_gl_proc_address_get() function as it is done in other
glue layers

2. provide all the extension functions in the EVAS_GL_API struct.

#1 approach is how it's done in other glue layers and the driver implementor can
provide new extensions easily.  It is however pretty annoying to get the
function prototypes right and use the function pointers and etc.

#2 approach provides all the extension functions in the struct so it's
definitely easier to use.  Adding new extensions can be a pain as people may
have to wait for new version releases.

For now, we thought it was OK to just throw them in the struct as in
#2 and try it out.  So, I've implemented this approach.  As for the extensions,
I've basically included all the extensions in gl2ext.h as EvasGL currently
provides all the GLES 2.0 functions.   In order to display the right
information, I had to override glGetString() with GL_EXTENSIONS as parameter to properly
display the supported extensions.

Also, I've added a few EGL extensions that have been
modified for EvasGL use.  For example, eglCreateImage/eglDestroyImage has been
defined as folllows.

   EvasGLImage  (*evasglCreateImage) (int target, void* buffer, int*
                                         attrib_list); void
                                            (*evasglDestroyImage)
(EvasGLImage image);

const char *evas_gl_string_query() function was added to return a string of
supported EvasGL extensions.  So essentially, a user can search this string to see
if the desired extension is supported.  if it is, he can use the functions.  He can
always check if the function pointers are NULL as well.

Take a look at the pach and let me know what you think.

______________

While I was adding the extension code, I've added a few fixes/ changes
to the EvasGL.

1. glDeletBuffers bug
- When I wad destroying evasgl context, I was deleting the context FBO with
glDeleteBuffers instead of glDeleteFramebuffers. This code in effect was
deleting BOs in other contexts and we had some funky behaviors as a
result.  The
bug has been fixed.

2. make_current
- I've made some changes to the make current code and also added a resource
context to the engine data.  the resource context is used for creating surface
texture/ fbos when surface/ context is created.  Before, i was using evas'
context but thought it'd be a good idea to use a separate context.

SVN revision: 64139

12 years agoefreet: typos
Boris Faure [Mon, 17 Oct 2011 20:29:25 +0000 (20:29 +0000)]
efreet: typos

SVN revision: 64137

12 years agoEvas font: Cache Harfbuzz font structure for fonts we use.
Tom Hacohen [Mon, 17 Oct 2011 14:24:49 +0000 (14:24 +0000)]
Evas font: Cache Harfbuzz font structure for fonts we use.

SVN revision: 64136

12 years agoEvas textblock: Beautify the DRAW_FORMAT_DASHED macro a bit.
Tom Hacohen [Mon, 17 Oct 2011 12:18:57 +0000 (12:18 +0000)]
Evas textblock: Beautify the DRAW_FORMAT_DASHED macro a bit.

Patch by Shilpa Singh.

SVN revision: 64135

12 years agoRevert "evas: Deal with harfbuzz API breakage"
Tom Hacohen [Mon, 17 Oct 2011 12:18:53 +0000 (12:18 +0000)]
Revert "evas: Deal with harfbuzz API breakage"

Sorry mate, but they broke API without bumping version, that's why I
didn't do this myself. You should probably add your version of harfbuzz.

This reverts commit 64057.

SVN revision: 64134

12 years agoEvas textblock: Cleaned up DRAW_FORMAT/_DASHED macros.
Tom Hacohen [Mon, 17 Oct 2011 10:43:42 +0000 (10:43 +0000)]
Evas textblock: Cleaned up DRAW_FORMAT/_DASHED macros.

Fixed compiler warnings.

SVN revision: 64132

12 years agoEvas textblock: Added dashed underline support.
Tom Hacohen [Mon, 17 Oct 2011 10:43:08 +0000 (10:43 +0000)]
Evas textblock: Added dashed underline support.

Patch by Shilpa Singh

SVN revision: 64131

12 years agomatch changes in ews code
Carsten Haitzler [Mon, 17 Oct 2011 07:00:38 +0000 (07:00 +0000)]
match changes in ews code

SVN revision: 64126

12 years agohandle buffer allocation failure - null buffer.
Carsten Haitzler [Mon, 17 Oct 2011 06:59:08 +0000 (06:59 +0000)]
handle buffer allocation failure - null buffer.

SVN revision: 64125

12 years agoEcore: fix maintainer-clean rule
Vincent Torri [Mon, 17 Oct 2011 06:56:04 +0000 (06:56 +0000)]
Ecore: fix maintainer-clean rule

SVN revision: 64124

12 years agosorry. we can't depend on svnrev in versions. wish we could. we can't :(
Carsten Haitzler [Mon, 17 Oct 2011 05:22:08 +0000 (05:22 +0000)]
sorry. we can't depend on svnrev in versions. wish we could. we can't :(

SVN revision: 64121

12 years agoFrom: "Sung W. Park" <sungwoo@gmail.com>
Sung W. Park [Mon, 17 Oct 2011 05:06:01 +0000 (05:06 +0000)]
From: "Sung W. Park" <sungwoo@gmail.com>

"... I've created a resource context per thread using TLS.  Since there is no
TLS support in Eina, I've added 4 APIs for that as well.  Another patch has
been submitted but i'll just include it in here as well. ..."

SVN revision: 64120

12 years agorequire recent eina
Mike Blumenkrantz [Sun, 16 Oct 2011 15:45:28 +0000 (15:45 +0000)]
require recent eina

SVN revision: 64115

12 years agoecore/animator - never be reached there
ChunEon Park [Sun, 16 Oct 2011 04:47:42 +0000 (04:47 +0000)]
ecore/animator - never be reached there

SVN revision: 64101

12 years agoeina: add eina_strbuf_trim/rtrim/ltrim()
Boris Faure [Sat, 15 Oct 2011 12:56:27 +0000 (12:56 +0000)]
eina: add eina_strbuf_trim/rtrim/ltrim()

SVN revision: 64098

12 years agoEvas: Check for dlsym and disable gl engines if not found
Youness Alaoui [Sat, 15 Oct 2011 09:31:04 +0000 (09:31 +0000)]
Evas: Check for dlsym and disable gl engines if not found

SVN revision: 64091

12 years agoEcore: IN6_IS_ADDR_V4MAPPED needs an in6_addr as input on OSX
Youness Alaoui [Sat, 15 Oct 2011 09:30:58 +0000 (09:30 +0000)]
Ecore: IN6_IS_ADDR_V4MAPPED needs an in6_addr as input on OSX

SVN revision: 64090

12 years agoEvas: Remove a useless ifdef that was causing dead code.
Youness Alaoui [Sat, 15 Oct 2011 09:30:33 +0000 (09:30 +0000)]
Evas: Remove a useless ifdef that was causing dead code.
You can't compile a gl_common .c file based on whether or not the SDL
header was included. The .c file will result in only one .o and since
the Evas_Engine_Sdl.h is not included by evas_gl_context.c itself, then
that ifdef will never be true.
gl_common should request a callback function pointer from the evas engine
for doing symbol resolution. This needs a refactor.

SVN revision: 64086

12 years agocommit changes and notes for matrchnig multitouch xcb stuff
Carsten Haitzler [Sat, 15 Oct 2011 03:34:01 +0000 (03:34 +0000)]
commit changes and notes for matrchnig multitouch xcb stuff

SVN revision: 64082

12 years agofor gustavo -> ews matching changes.
Carsten Haitzler [Sat, 15 Oct 2011 03:19:25 +0000 (03:19 +0000)]
for gustavo -> ews matching changes.

SVN revision: 64081

12 years agoEvas: Fix compiler warnings wrt uninitialized variables.
Christopher Michael [Sat, 15 Oct 2011 01:02:55 +0000 (01:02 +0000)]
Evas: Fix compiler warnings wrt uninitialized variables.

SVN revision: 64079

12 years agoand now make my modifications to multittouch this morning also work
Carsten Haitzler [Fri, 14 Oct 2011 09:11:49 +0000 (09:11 +0000)]
and now make my modifications to multittouch this morning also work
again with mpx automatically. yay!

SVN revision: 64071

12 years ago1. make inline image ee buffer non-alpha by default. thats the default
Carsten Haitzler [Fri, 14 Oct 2011 07:28:45 +0000 (07:28 +0000)]
1. make inline image ee buffer non-alpha by default. thats the default
for all eee's. this was wrong.
2. make sure it sets back data when finished - check data in and check
it out.

SVN revision: 64069

12 years agoput in hardcoded bitmask define if headers dont provide it.
Carsten Haitzler [Fri, 14 Oct 2011 07:23:09 +0000 (07:23 +0000)]
put in hardcoded bitmask define if headers dont provide it.

SVN revision: 64068

12 years agoemotion: work with older version of gstreamer.
Cedric BAIL [Fri, 14 Oct 2011 04:01:43 +0000 (04:01 +0000)]
emotion: work with older version of gstreamer.

SVN revision: 64058