Previously, due to propagation to parent, an event could have
been received more than once by an object. This triggered
strange behaviour in edje for example where you could receive
mouse,down,1 signal many time for one swallowed object.
This patch is a fix for that problem, I hope it doesn't break
anything (e17 and elementary_test run fine here, but report
any break related to events please).
SVN revision: 46869
Canvas was (ab)using the same callback signature as Objects, so you
always got a confusing NULL parameter.
Just clean it up to be Evas_Object_Event_Cb and Evas_Event_Cb, each
with its own signature.
SVN revision: 46206
- they dont. in reality.
2. major text rendeering speedups. up to 41% in textblock intl, 33% in
textblock basic, 12-20$ in other text rendering tests. generic eina hash's
are just tooo slow for what we are doing there. specialised "Fash"
blocked-array.
3. still LOTS of optimisations left.
SVN revision: 45829
1. the concept of callbacks for a canvas as a whole. add/ del/ del_full these
2. focus in+out events for the canvas as a whole - can help solve some issues
with inoput methods + ecore-imf + entries (like edje_entry)
3. add callabcks to be called before/after flush of display.
SVN revision: 45761
This patch adds some stuff for smart callback description/instropection, which
is still untested but doesn't break anything that's out there now. Should help
with bindings later on.
Also some parenting guidelines for smart objects, so it's easier to spawn a
subclass out of another. Look at Box and Table for an example on this.
And again, rebuild everything that uses smart objects after this update, or
the world will turn into a happy place where lawyers are no longer needed...
and we don't want to upset the lawyers.
SVN revision: 45043
Evas image load was always reporint "generic" error, since it was
disconnected from actual loader modules.
This commit will break the module loader API (as it's restricted to
inside Evas, this should be no problem). The return was turned into
"Eina_Bool" for clarity, while an extra "int *error" is responsible to
report errors. This approach was choosen to force compiler warnings
and to try avoid mistakes as EINA_FALSE == EVAS_LOAD_ERROR_NONE and
thus we'd get opposite behavior if something slips.
Most loaders play well, except by eet that does not provide means to
know if the file open failed due missing file, incorrect format or
corrupted file :-(
Please report any issues. I added eina_log debugging to loader
functions, just run your Evas application as:
EINA_LOG_LEVELS=evas_main:4 your_app
SVN revision: 44666
This code should be cleaner and easier to understand. It also provides
the ability to spread image decompression on all CPU core. I currently
set it to the exact number of CPU core you have in your machine, if you
find case where it slow down your EFL apps too much, we can reduce this
to give at least one core to evas.
All previous bugs related with async preload are gone, hopefully no
new one are in. Please report any problem with backtrace to me.
SVN revision: 44537
changed. still renders all, but better now. keeps map surfacer around for
shits and giggles until map unset or object deleted. als be able to set
smooth map and disable alpha (for smart objects)
SVN revision: 43362
change evas_map to return a structure that serves as an array of
points. This way we'll know for sure the number of points in it. Right
now it's hardcoded to 4, so check it, but in future we can just allow
more points and it should work.
added docs. I'm not sure about most of it, so it would be good to have
someone to review and fill in more, maybe that's raster? Grep for
"TODO" and you'll see the missing stuff. It would be good to add
examples in evas_map_point_coord_set() and
evas_map_point_image_uv_set()
SVN revision: 43211
Hey raster,
Here is the non intrusive patch I talked to you about. Please apply it as it
introduces some fixes, some improvements and mostly and underlying
infrastructure for future RTL improvements.
(note hebrew & yiddish seem fine, but things expedite test seems to show are
wrong (why i don't know as i dont speak the langs- just comparing to pango /
gtk output):
arabic seems lsightl wrong (maybe composition chars not working?)
gujarati - also seems wrong
malayam - also looks wrong
persian - looks wrong
sinhala - looks wrong
tamil - looks wrong
these are what, appear to me, to look wrong. why they look wrong, i don't
know. i'm guessing its compositiong not being handled. but i dont's peak,
read or write any of these languages so i am unsure of what it really should
be like, why and how to fix it.
anyone want to put up a hand? (everything else is displaying fine as best i
can tell - the langauges i read/speak/somewhat understand are working fine).
SVN revision: 42814
be way too big to ever allocate. probably code can do with other fixes too.
also make jpeg loader rudametarily understand load regions. very brute-force.
but enough for just this moment to do testing.
SVN revision: 42507
Improvements: Now evas rendering loop is the one responsible to
initialize the surface to 0 correctly (taking into account surface
alpha and object opacity). This will reduce the number of memset
we do.
Note: Current software_x11 (xlib and xcb) are buggy. They are
copying too much data when the surface use a mask. That's why
two memset are left in their code. They could be removed, but
we should fix the surface we copy on change (look at mxob user
and evas_software_xlib_x_output_buffer_paste).
SVN revision: 41206
also rename evas_layer_free() to evas_layer_free_objects() as what it
do now, make _evas_layer_free() as static and use it both cases.
SVN revision: 41123
the parent.
When we insert object inside a smart object, they could be attached to
another layer. As long as ref counting work, nothing wrong will happen.
But during destruction of an Evas, we were just looping over all layers,
destroying each of them, without checking for refcounting. This could
cause SEGV.
This patch introduce a third loop for wiping out all layers after
destroying all Evas_Object. So no more SEGV, and no performance
regression.
Note: Do not rely on evas_object_layer_get on smart object's child, it
could give you the wrong answer.
SVN revision: 41046
none (ie default) and the engines actually understand it and use it.
2. fixes to scalecache and cserver too. more toto's done and its now been
stress tested by me - and i think cserve is ready to go gold. just enable it
with export EVAS_CSERVE=1 in your env for any eflapps - and run evas_cserve
(cmd-line options avalable plus cmd-line tol to query settings change on the
fly and query statsitics and state)
SVN revision: 40536
is it ok?
1. it can be --disabled in evas's configure, but i think it works WITHOUT
disabling it (runtime) as it falls back to the old way of loading
2. it may cause build problems on some platforms - without it being enabled
we won't find out, so enable.
3. it needs enabling runtime to make use of it so it should be safe for now
until you enable it.
what is it?
it is a SHARED cache server - that means images loaded are loaded BY the
cache server (not by the actual process using evas). images are shared via
shared memory segments (shm_open + mmap). this means only 1 copy is in all
ram at any time - no matter how many processes need it , and its only loaded
once. also if another app has already loaded the same data - and its in the
cache or active hash, then another process needing the same stuff will avoid
the loads as it will just get instant replies from the cache of "image already
there". as it runs in its own process it can also time-out images from the
cache too.
right now you enable it by doing 2 things
1. run evas_cserve (it has cmd-line options to configure cache etc.
2. export EVAS_CSERVE=1 (im the environment of apps that should use the cache
server).
it works (for me) without crashes or problems. except for the following:
1. preloading doesnt work so its disabled if cserve is enabled. thisis
because the load threads interfere withthe unix comms socket causing
problems. this need to really change and have the cserve know about/do
preload and let the select() on the evas async events fd listen for the
unsolicited reply "load done". but it's not broken - simple preloads are
syncronous and forced if cserve is enabled (at build time).
2. if cserve is killed/crashes every app using it will have a bad day. baaad
day. so dont do it. also cserve may be vulnerable to apps crashing on it - it
may also exit with sigpipe. this needs fixing.
3. if the apps load using relative paths - this will break as it doesnt
account for the CWD of the client currently. will be fixed.
4. no way to change cache config runtime (yet)
5. no way to get internal cache state (yet).
6. if cache server exist - it wont clean up the shmem file nodes in /dev/shm
- it will clean on restart (remove the old junk). this needs fixing.
if you fine other issues - let me know.
things for the future:
1. now its a separate server.. the server could do async http etc. loads too
2. as a server it could monitor history of usage of files and images and
auto-pre-load files it knows historically are loaded then whose data is
immediately accessed.
3. the same infra could be used to share font loads (freetype and/or
fontconfig data).
4. ultimately being able to share rendered font glyphs will help a lot too.
5. it could, on its own, monitor "free memory" and when free memory runs
load, reduce cache size dynamically. (improving low memory situations).
6. it should get a gui to query cache state/contents and display visually.
this would be awesome to have a list of thumbnails that show whats in the
cache, how many referencesa they have, last active timestamps etc.
blah blah.
please let me know if the build is broken asap though as i will vanish
offline for a bit in about 24hrs...
SVN revision: 40478
we change cur.geometry in the code (did a grep to locate this). I hope I did
spot all users, as I didn't see bug in exec_buf, efm and in window title, I
am confident enought to break svn again.
* WARNING * This change can cause visual bug. Please report.
SVN revision: 40039
why it was disable the first time, so it could lead to some graphic bug.
Please report any strange behaviour.
*WARNING* This could really introduce some visual bug.
SVN revision: 39940
enabled. the blending is not working/complete. the neon for fills and copies
isnt actually faster though currently :(
2. scalecache infra - disabled for now. working on it.
SVN revision: 39723
* evas_engine_info_set() returns now an int, to inform if
an error occured or not when setting the info of the engine.
* in the Evas_Func structure, the setup() method returns an int
* all the engines are updated
I'll fix ecore_evas and ewl later (the compilation is still fine).
Gustavo: should I add EINA_WARN_UNUSED_RESULT at the end of the
evas_engine_info_set() function ?
SVN revision: 39670
let's stop replicating these macros over and over again, also flag
evas functions with attributes to help with optimizations.
TODO:
* move functions returning "int" as boolean to Eina_Bool
* move api entry to EINA_SAFETY_*
* document api
SVN revision: 39598
* evas/src/lib/engines/common/evas_font.h,
* evas/src/lib/engines/common/evas_font_draw.c,
* evas/src/lib/engines/common/evas_font_load.c,
* evas/src/lib/engines/common/evas_font_query.c: Add cache for font kerning.
This patch give something around 2% for all tests around text in expedite,
except for Textblock Intl where it give a 3 times boost.
Regarding text rendering speed, something is strange when used by evas_pipe.
All tests using Styles are around 40% faster without evas_pipe. 30% faster
for Text Change. But Text Basic 7% slower. So it should be possible to have
faster rendering when using evas_pipe for font rendering.
SVN revision: 38993
1 - use inlist as regular list uses non-thread safe mempool;
2 - lock around image loading, so if main thread requests pixels right
before worker thread is loading them, you don't get ie->info.module
to NULL while it would be used (triggered from engines/common).
Maybe this should be handled by a global mutex elsewhere instead of
per-image mutex, but it has more granularity now.
3 - emit "preloaded" callback if it was canceled to be loaded from main
thread.
Please someone review these changes.
SVN revision: 38312
if image was already preloaded, inform user.
regular use case is to have image hidden, ask for preload and then
show image on callback, if there is no callback, image is never shown.
SVN revision: 38236
tyo fix up some of these breaks first and there isn't a lot of time devoted
to this. so revert this. it's in svn history so we can dig it out any time we
like.
SVN revision: 37453
1. nearest scaling is now broken - it's always linear interpolation. this
will lead to slowdowns. i need to fix this - a must.
2. i think it's time i put in a transformed image cache that can cache an
image object at a transform (and share it) automatically.
3. transforms in non-software-engines will not work - broken. need to at
least do xrender and gl engines.
any volunteers to help?
SVN revision: 37447
Draw back: When we are destroying an Evas canvas, we loose all cached font
that are not used anymore.
A correct fix would be to link Fndat to the Evas that provide and use them.
And only delete them when no more Evas reference them.
SVN revision: 37353
that's about it. a bit hacky - but works and frankly.. the idea is that u'd
set a scale factor once really and not change it per obj... most likely.
SVN revision: 35896
Some people is using it for some time now without problems, so I'm
adding it to SVN to get some broader use. Remember to recompile ALL
libraries that depend on Evas as it will change the
EVAS_SMART_CLASS_VERSION and old classes will fail to load.
This will also change Edje so it will postpone _edje_recalc() to
render time, calculate() callback, however some methods will force
early recalculation.
SVN revision: 35860
- some enignes break as they dont have the stubbed out functions, and
xrender/gl engines dont even implement the drawing and need to (but are
stubbed out).
SVN revision: 35677
Image_Entry flag structure. This fix a bug with 16 bpp software engine.
* Change image loader module API to take any Image_Entry. Same goes
for evas_common_image_premul and evas_common_image_set_alpha_sparse.
* Use new eet API: eet_data_image_read_to_surface.
SVN revision: 34728
This patch fixes the problem with bitfield of signed types (ie: char),
where the bit would be used for the signal, so 1 is considered -0 and
thus 0.
Move all the single bit fields to Evas_Bool, making it clear and also
avoiding these problems since Evas_Bool is unsigned char.
SVN revision: 34631
By having a layer as a short (16 bits) we can pack it together with
the bitfields, saving 4 bytes per sub-struct, 8 bytes in total, also
bringing the struct down from 4 to 3 cachelines on my laptop.
Rationale: layers are mostly used to differentiate groups of objects
and they stacking, usually we have few layers and we use very large or
very small numbers to make a layer be at the top or at the bottom, but
usually we don't need so many layers.
Caution: code that use values like 999999 will break, so fix your
code! I'll provide another patch to fix all the CVS using these large
values.
SVN revision: 34420
This saves 20 bytes, bringing Evas_Object to 200 bytes, by moving data
specific to smart objects to their own struct (Evas_Object_Smart).
There is still one remaining member that could be removed:
smart.smart, this is used mainly to identify if one object is a smart
object or not. One possibility would be to add a bitfield to state
that, but another possibility is to check Evas_Object::object_data
and see if it's a smart or not.
SVN revision: 34419
Interpolation color_space (now ASHV or ARGB) was being used inside a
struct with 4 byte alignment. Remove it from the struct and make it a
bitfield so can be packed with the other fields. This saves 2
integers, so 8 bytes.
SVN revision: 34418
This is a repack of bitfield members, was tested on GNU/Linux + GCC 4.1.2
and works fine. Needs further testing on other compilers.
SVN revision: 34417
Size hints are useful, but wasting 36 bytes for it on every object is a bit
too much: clippers and lots of other objects will have no need for it.
Now it's a pointer to a struct that will be allocated just when some value
is set, wasting 4/8 bytes more for the pointer when it is used, but saving
32/28 bytes when it is not.
This will also help to have alignment properties in future, that can come
as hints, without too much impact on memory consumption.
SVN revision: 34412