This make it possible to use the object tree to reduce the number of object, we need to explore to know
what is under a specific position. First used by propagation event code. That code is now 4 times faster, enjoy !
As a side cost evas_object_move goes from 925 to 980 valgrind cycle on my computer, so not something you
will notice.
NOTE: if you notice any breakage regarding event propagation, map, cats, minor or major, please report to
me ! I hope I didn't loose my mojo, with such a scary change, I have a big chance to get it back !
SVN revision: 70564
NOTE: This should be part of evas_render itself and not
delegated to the engine. So cleaning things to make it easier
during evas_render rewrite.
SVN revision: 70503
Note: this two were broken. Current plan to bring
that feature back in, is to attach this information
to Evas_Text_Prop during the prepare stage. This
would improve both single and multiple core rendering
without increasing the number of lock and the complexity
of the code.
SVN revision: 70501
NOTE: other things that may join it in the near futur EVAS_FRAME_QUEUE,
EVAS_METRIC_CACHE and maybe EVAS_WORD_CACHE also. This is all part of
cleaning up our rendering path so we can actually improve it more easily.
SVN revision: 70499
NOTE: some of this function should be moved as inline, but that's to late for a change
I think. So we will fix that if needed.
Second point, I am not happy with is eina_inarray_insert and eina_inarray_insert_at. The
naming is really poor.
SVN revision: 70352
Attached patch fixes a bunch of warnings in evas doc:
- Make remaining EINA_{TRUE,FALSE} @c, removing some undef link in same
time
- Make remaining NULL @c
- Fix the color space list
- small random fix
SVN revision: 69956
Due typo the weight was being handled as an integer, not floating
point. It worked with examples since they were usually being round to
int after being sum (0.3 + 0.7 -> 1.0, 3 + 7 -> 10).
SVN revision: 69910
There is a reference to evas-load-error-str.c in
the evas_load_error_str doc, but this file doesn't exist
anymore so I made the doc link to evas-images.c which use
several times this function.
SVN revision: 69754
Subject: [E-devel] [patch] evas doxygen doc (1)
There are a lot of small things to fix in the evas doc. Here is a first
batch. Fixes are trivial and obvious. Just a question about the first
hunk, I removed the <tdb> that replaced valid email for companies
because it makes doxygen complain about the unknown tag and, well, it
doesn't give any information either. Was there an explanation about
adding this ?
Joined is the patch along with the doc build log diff.
SVN revision: 69749
This works in linux and windows, and should fix shm_detection on BSD (including Mac)
BSD, Mac and solaris users : please check that it compiles and shm_open is detected
SVN revision: 69612
Subject: image data get/set pairing issue
I found a bug about pairing
evas_object_image_data_get/set(eng_image_data_get/put).
It was added to count checked_out for paring
eglMapImageSEC/eglUnmapImageSEC.
In case of calling evas_object_image_data_set() twice after calling
evas_object_image_data_get(), dyn.checked_out has -1.
Then, if evas_object_image_data_get() and evas_object_image_data_set()
is call, it can't call eglUnmapImageSEC().
If dyn.checked_out has minus, it can make some problem.
So, I fixed this problem.
Please find enclosed patch file and let me know if I misunderstood.
SVN revision: 68504
From: Thanatermesis <thanatermesis.ecvs@gmail.com>
Subject: [E-devel] LDFLAGS with -Wl,-z,defs
Aparently if you add the option "-Wl,-z,defs" to your LDFLAGS, there's some
libs that doesn't compile, like evas and e_dbus, there's some logs:
SVN revision: 68464
subject: [E-devel] [Patch] Add override gl apis for osmesa
When an application use glBindFramebuffer or glBindRenderbuffer via
evas_gl after loding libosmesa.so,it shows segment fault.
Because glBindFramebuffer and glBindRenderbuffer are not overrided.
So, I fixed it.
SVN revision: 68391
NOTE: I would like to do the same with software SDL 16bits engine.
But as we don't have a buffer_16 backend, I am likely to just remove
it and use buffer conversion code to match a 16bits target. This
will come with a performance impact, that will make it useless. So
I am just tempted to completly remove it.
SVN revision: 68352
Subject: [E-devel] [patch] evas - preventing retard mouse event
process in evas_object_event_callback_call
I made a small patch to prevent retard mouse event process.
At certain circumstance (like as genlist select callback + naviframe
item push),
some events are repeat processed.
If some evas_objects're iterating in evas_event_feed_mouse_up and
mouse_out event's emitted by other interrupt(such as naviframe item
push),
then some evas_objects events are multiple processed by
evas_object_event_callback_call.
More elaborating it with a example.
There are a genlist and a multibuttonentry on genlist item.
When a user clicks multibuttonentry then evas will process mouse down
and up.
in evas_event_feed_mouse_up, it gets evas object list to process mouse
events.
Then in the evas object list, there are two evas objects - rect and
textblock.
Two objects have its own parents.
the rect has below parents.
----------------------------------------
edje - genlist item
elm_genlist_pan
edje
multibuttonentry
box
entry
els_scroller (0x2a601788)
rect <== the rect
the textblock has below parents.
----------------------------------------------
edje - genlist item
elm_genlist_pan
edje
multibuttonentry
box
entry
els_scroller(0x2a601788)
edje
elm_pan
edje
textblock <== the textblock
(note : two evas object have same parent (els_scroller))
So normally mouse up callbacks event propagates to its own parent.
the rect is processed to genlist item. and the textblock is processed
to genlist item.
but when els_scroller is processed, it's blocked by checking event
type and event id checking.
Mouse Up(rect) -> Mouse Up(textblock)
event_id (3) -> event_id (3)
evas_object_event_callback_call(Evas_Object *obj, Evas_Callback_Type
type, void *event_info, int event_id)
{
...
if ((obj->delete_me) || (!obj->layer)) return;
if ((obj->last_event == event_id) &&
(obj->last_event_type == type)) return;
<=== blocked
However if naviframe item is pushed in the middle
of mouse up processing.
It can break into mouse up. So it's processed like below.
Mouse Up(rect) -> Mouse Out(rect) -> Mouse Out(textblock) -> Mouse
Up(textblock)
event_id (3) -> event_id(4) -> event_id(4)
-> event_id(3)
(note Mouse_Out is made by naviframe item push for event freezing)
If that, there's no mechanism to block that repeat processing same
event.
So I suggest this patch.
This patch blocks old events if there's no reason to process.
(It blocks old mouse_up event because mouse_out is processed.)
And I think it also clear the bug in
"[E-devel] event repetition with elm_naviframe/elm_genlist"
SVN revision: 67879
if file was chaned by somebody, it was added to dirty list during cache request.
currently this dirty image added to cache->dirty list and never freed until image shutdown.
but dirty image of chaned file never used so add delete code for memory efficiency.
and fix bad indentation.
SVN revision: 67604
void pointer causing a segfault. The attached patch fixes the issue
and allows an expedite run to complete.
By: Will Newton <will.newton@gmail.com>
SVN revision: 67543
option.
It should have been OR instead of AND operator.
When the image object alpha is on "OR" the rotation angle
is not "0", direct rendering isn't allowed. However,
allow direct rendering if EVAS_GL_DIRECT_OVERRIDE=1 is set.
SVN revision: 67521
This optimization is significant for rendering to a large surface
because it'l save an extra copy overhead as well as an extra rendering pass.
To enable it, you can give EVAS_GL_OPTIONS_DIRECT hint in the surface
config options_bits. The following conditions have to be met in order
for evas to render directly into the Evas' window. If they are not met, the
engine will fallback to rendering to an FBO as it normally does.
conditions:
1.) All the GL calls have to be called using the pixel_get_callback function.
This is necessary for the evas object order to be maintained.
2.) Alpha must be disabled on the image ojbect that renders evas_gl.
3.) No rotation allowed.
One way to override above condition is to set EVAS_GL_DIRECT_OVERRIDE=1 but
there is no guarantee in its behavior.
Currently, this optimization is added for gl_x11 engine only.
SVN revision: 67388
Make the functions static as well to avoid gcc warnings like this:
warning: 'ALPHA_SSE3' is static but used in inline function 'sub4_alpha_sse3' which is not static
Signed-off-by: Stefan Schmidt <stefan@datenfreihafen.org>
SVN revision: 67355
This really just filters them out. The solution is not complete, nor is
it the best one. But this fixes the bugs for the meanwhile.
SVN revision: 67327
We only check the value of references if EVAS_CSERVE isn't defined, so
no need to define it or assign it if EVAS_CSERVE isn't defined.
SVN revision: 67304
be 1.1 (or 1.5) for the last release. too late. THIS is why i'm sick
and tired of all the bloody separate libs that have to be versiioned
and build and released separately. :( too many places to go fix up per
release.
SVN revision: 67284
found was causing compiler optimizations to, in some compielr versions
and optimization levels, to skip a func in the eva stest suites...
which is wrong. no more pure mumbo. gone entirely.
evas_textblock_string_escape_get() was the one.
SVN revision: 67266
if the eng setup has a NULL surface, and if the RenderEngine has an
existing surface, that means we are hiding/closing the window, and
thus should free the existing RenderEngine Window.
SVN revision: 67160
to ensure backward compatibility. Previously, the user
simply declared a Evas_GL_Config object but this can
cause problems if more config options are added. So,
we have Evas allocate the config object for the user
so it can handle addition in the future.
Also, added some safety code around _extensions_init
SVN revision: 67141
old version of this w/ the 3 includes Should be working, but I've
tested it on 2 machines now, and it fails on both with those lines in
there, so I am removing them).
Make wayland_egl engine Actually work and draw stuff now (too many
code changes to list them all separately). See http://i.imgur.com/i2eBE.png.
SVN revision: 67128
On macosx i386, that code fails because even though __VEC__ is defined,
the compiler doesn't understand the 'vector' keyword (that macro is
irrelevent here). So there was no way to make evas compile for ppc if
altivec was not supported by the compiler.
SVN revision: 66966
* This feature requires libOSMesa to be installed.
One caveat with OSMesa is that a surface config (ie. Depth Buffer and etc)
is associated with a context rather than a surface, which is the case
in EvasGL. So for now, when a user specifies a surface config, it gets
associated with the first context that the surface does a make current to.
For typical usage case, this shouldn't be a prolem. Will need to fix it
eventually.
SVN revision: 66896
Handle 'moving' nad 'resizing' of objects if they are not 'frame objects'
Add code for frame_object_get/set functions.
Fix some formatting.
SVN revision: 66536
Subject: Drawing objects by pixman
* Extend pixman support to allow other operations to use
pixman when doing software rendering. On x86 this isn't useful
but on ARMv7 with NEON pixman happens to do better with image
blending and nearest scale blending.
* Add tiled rotator for 32bit display as an option.
SVN revision: 66478
intel drivers don;'t like it for some odd reason - i'm trying to track
it down but i can't sanely try middlegrounds right now (eg dont
dlopen/dlsym but actually directly assign symbols etc.), so back out
and let's figure this out before it goes back in :(
SVN revision: 66308
1. Make Obj replacement and Par Sep less confusing.
2. We'll may, at some point, use the Unicode NewLine char instead of \n.
so it's now easily replaceable.
SVN revision: 66255
currently evas_object_event_callback_call checks _evas_event_counter
for preventing object's callback called several times in one evas event.
but it use global variable(_evas_event_counter), it can be changed while
procssing same event.
for example , evas_event_feed_mouse_up.
If there are several object in e->pointer.object.in and object 1's callback
create new evas event, object 2 cannot now event id.
so I change callback call api, and object callbacks can decide wheather it deal with that event.
SVN revision: 66234
addition to evas gl backend.
Working on different mobile devices, we've noticed that the cost of context
switch (MakeCurrent) in GL can be very expensive depending on the driver
implementation. To minimize the poorly written driver's context switch
overhead, I've implemented a state tracking layer on top of the driver
implemented GL.
Essentially, this layer wraps all the GL/Glue(GLX/EGL) APIs and manages its own
state changes. Internally, only one real GL context is created and logical
contexts are created every time a user requests context creation. The logical
contexts keep track of its states and sets only the necessary states
(the ones that are different than the current ones)
when there is a MakeCurrent request. The real MakeCurrent gets called when
there is a Surface/Window change request.
The GL library is dlopened and all the APIs are dlsym'ed and wrapped
accordingly. All the GL functions are in evas_gl_core.{h,c}.
Here's a very simply flow of the code.
- all the APIs are exported as function pointers (*glsym_glBegin),
(*glsm_eglCreatContext), and etc.
- all the native GL/Glue(GLX/EGL) APIs are dlsym'ed as _sym_glBegin,
_sym_eglCreateContext, and etc.
- all the fastpath APIs are implmemnted as fpgl_glBegin,
fpgl_eglCreateContext, and etc.
- if faspath is seletected, the exported APIs are set accordingly
ie. glsym_glBegin = fpgl_glBegin;
- default mode is the regular gl symbols are directly set.
ie. glsym_glBegin = _sym_glBegin;
I have an Environment variable where you can set it to three different Modes
EVAS_GL_FASTPATH = 0 // Default mode. Regular GL symbols are directly loaded
EVAS_GL_FASTPATH = 1 // Fastpath mode. Takes the path described above.
EVAS_GL_FASTPATH = 2 // Wrapped mode. All the regular GL functions are
wrapped once. This can be used for various
purposes
Since all the GL symbols are now loaded in this library, I took out all
the gl symbol loading parts in the evas gl backend as you'll see in the patch.
The changes to the engine and the backend itself is pretty minor.
There are still some known issues to hammer out but I thought we're at a good
place for an initial version so that my source doesn't diverge too much.
Known Issues and To Do's
* Current GL Fastpath version doesn't support multiple threads. Instead of
having one global real context, I would need to do it for each thread. I'll
get on this soon.
* Issues running Evas GL on certain conditions. When running the elementary
test (with gl engine), if you run ELMGLview test that runs in ON_DEMAND mode,
everything works fine. BUT, when you run the ELMGLView test in ALWAYS
mode, the subsequent elm tests shows blank screen. When you destroy the
GLView window, everything else comes on fine.
* Resource protection code. This actually applies to Evas GL code in general
as well. Since all the resources are shared among all the contexts that get
created, I would like to eventually have a resource protecting mechanism that
prevents access to resources outside of its context unless specifically
specified.
I'm attaching three files
- evas_gl_core.h, evas_gl_core.c, fastpath.patch
To get the code running...
- copy evas_gl_core.{c,h} to src/modules/engine/gl_common/
- apply the fastpath.patch
- compile/install evas
- to run with fastpath GL (ie. % EVAS_GL_FASTPATH=1 ./evasgl_sample1)
SVN revision: 65891
Subject: Fix Neon build with Thumb-2
In assembly part of function evas_common_convert_rgba_to_32bpp_rgb_8888_rot_90:
Don't use 3-operand add instructions (e.g. add r2, r5, #2) as this is
not supported in unified syntax.
SVN revision: 65768
tweet: hey, SeoZ here, I really like Microsoft. I mean REALLY like them. I think they're a fantastic company and I hope everyone uses winmo!
SVN revision: 65469
quickly when uf your-encoded anything due to using IFAST decoder (and
encoder). this does drop speed for decode and encode (except encoding
< 60% quality where it now uses IFAST), but we don't see progressively
worse artifacts.
SVN revision: 65293
in most cases, it is performed twice inside and outside of the function in inefficient way.
and calling of events_frozen in the passes_through() is not understable in the view of functional consistency.
Check is separately would be better.
SVN revision: 65269