defined in GL/glext.h (which is not included if we are building for
SGX). A build for SGX uses EGL which defines GL_WRITE_ONLY_OES instead.
NB: I am not entirely sure that this is the 'proper' fix, but it
allows evas to build again.
SVN revision: 76540
Firstly, I want to say hallo to everyone, it is my first message to
that
list.
I had prepared several patches for eina module. One of them fix
problem with
you are having highly discussion, checking if function arguments
are not
NULL.
What’s more:
- I had proposed patches for eina_tiler and
eina_rectangle
functions, now below zero values ( distance and coordinstes ) are
not
acceptable.
- Documentation for eina list specified, it should be
more clear
now and eina stringshare_strlen fixed (NULL had length 0 ).
- eina_convert_atofp wrong return value if fp is NULL
(EINA_TRUE
instead of EINA_FALSE)
I had also attached diff to AUTHORS, ChangeLog and NEWS.
SVN revision: 76498
This give another 10% speed improvements (and also reduce massively
memory allocation as we do it once for one array instead of thousand
of time per list of parameters).
SVN revision: 76433
This will give a major boost of performance for all insanely huge theme
you try to compile. Get down from 13s to 4s. Still some room for improvement.
SVN revision: 76427
Timeout messages are given a 408 HTTP status code (request timeout), instead
of a CRIT() message; this gives chance to the completed callback do whatever
is necessary when a timeout happens.
SVN revision: 76419
Thsi will also fail if the symbol is there but the init function just fails
to deliver EINA_TRUE. After to much digging and wondering why dlsymb and
dlopen don't behave Daniel was able to point out to me that this warning also
gets shown if init fails.
Adjust the warning message to avoid others spending to much time on this as well.
SVN revision: 76409
distros such as gentoo will rely on actual wanted dependencies and it
sucks to not have a way to forcefully enable or disable a dependency.
Scenario:
- user installed eio for some weird reason.
- user compiles edje, which was finding eio. Without user desire.
- user removes eio because it's not needed anymore.
- edje is broken.
Most systems will not have this problem because the package is built
on a pristine system where all dependencies are installed, then the
package is compiled and is impossible to have the user to do such
thing. Not on gentoo and others.
And this commit has a second benefit to disable eio until it's
properly fixed for edje_codegen :-)
SVN revision: 76361