Ouch, that was nasty! src/lib/cache/evas_cache_image.c was assuming
all Image_Entry were RGBA_Image (why?!?!?), thus doing the cast and
having mutexes and other to operate on garbage (possibly crashing due
segv). This happened to be the case with Soft16_Image used by
software_16 engines.
I'm not sure, but this may fix problems that people noticed with
async-render hanging their systems even if not explicitly enabled
during runtime.
I also found it quite strange the number of locks required by this
code! Clearly we could use macros to simplify and avoid bugs, but
maybe some atomic_inc/dec code should be used to remove half of the
mutexes in that code?
/me wonders what kinds of bug more we can expect from this code :-/
SVN revision: 50300
Software 16bpp variants were crashing if using preload as the
Image_Entry that was used by preload did not keep the loader and
module information, since these were only set in the
Soft16_Image::source. Now we copy these members while the source is
live.
SVN revision: 50299
Just set the EVAS_CPU_NO_NEON environment var if you want to disable Neon.
The same is done for other specific CPU codes, such as MMX, MMX2, SSE, etc.
SVN revision: 50271
The notnull.cocci script from Coccinelle finds places where you check if a
variable is NULL, but it's known not to be NULL. The check can be safely
removed. For example, this code would be caught by notnull:
if (!var) return;
if (var && var->fld) { ... }
It's needless to check again if var is not NULL because if it's in fact NULL,
it would have returned on the previous "if". This commit removes all the
trivial places where this pattern happens. Another patch will be generated for
the more complex cases.
SVN revision: 50241
Bug found by the "andand" semantic patch:
// The right argument of || or && is dereferencing something known to be NULL
//
// Confidence: High
// Copyright: (C) Gilles Muller, Julia Lawall, EMN, INRIA, DIKU. GPLv2.
// URL: http://coccinelle.lip6.fr/rules/andand.html
// Options:
@ expression@
expression E;
identifier fld;
@@
- !E &&
+ !E ||
<+...E->fld...+>
@ expression@
expression E;
identifier fld;
@@
- E ||
+ E &&
<+...E->fld...+>
SVN revision: 50118
The new wordcache option is an environment variable called:
EVAS_WORD_CACHE_MAX_WORDS
set this to a number between 1-500 to change the cache size.
Larger values (40+) fix regressions in a few of the expedite tests.
SVN revision: 50047
Code with side effects is NOT pure. Cdoe where you can ignore the return
value because you care about the side effects should not have WARN_UNUSED...
Please only add such things when you've checked
SVN revision: 49984
If you are having any problems with Arabic shaping please make sure you have FriBiDi version 0.19.2 installed at the time of compilation. If your distro doesn't have that version yet, make them upgrade, it's an year and a half old.
SVN revision: 49949
WARNING: I AM SORRY, BUT IF NO ONE IS PLANNING TO BECOME
THE EFFECTIVE MAINTAINER OF THAT PIECE OF CODE, I WILL
REMOVE IT NEXT WEEK.
SVN revision: 49945
The evas_debug functions are just called from MAGIC_DEBUG errors, that
are quite critical, so messages should come out by default, and do
that loud with a CRITICAL warning.
Raster, cedric and others: now that we have Eina as a common
dependency between Evas and Ecore, maybe it is worth to unify the
debug system on top of Eina logging? It already provides "abort on
critical", runtime configurable log levels. I'd say the same about the
magic and safety checks themselves (eina_magic and eina_safety).
SVN revision: 49643
Brett Nash
Cedric BAIL
Christopher Michael
Gustavo Sverzut Barbieri
Eduardo de Barros Lima
Lucas De Marchi
Carsten Haitzler
Jonathan Atton
Tom Hacohen
Vincent Torri
Bruno Dilly