Michel,
Is it now possible that drivers that still do a DRI lock/unlock in
render_start/render_finish will now do state validation with the
incorrect window size?
Is it necessary to have *two* checks for dirty state, before and after
the render_start?
Keith
On Sat, 2009-10-03 at 09:04 -0700,
On Mon, 2009-10-05 at 10:15 +0100, Keith Whitwell wrote:
Michel,
Is it now possible that drivers that still do a DRI lock/unlock in
render_start/render_finish will now do state validation with the
incorrect window size?
Hmm, possibly.
Is it necessary to have *two* checks for dirty state,
http://bugs.freedesktop.org/show_bug.cgi?id=24318
Summary: Fix building of GLSL demos which use M_PI when M_PI is
not defined by math.h
Product: Mesa
Version: git
Platform: x86 (IA32)
OS/Version: Cygwin
Status: NEW
http://bugs.freedesktop.org/show_bug.cgi?id=24217
--- Comment #2 from Andrey Tsyvarev tsyva...@ispras.ru 2009-10-05 06:18:44
PST ---
In the situations that I have checked, the problem with copying the texture did
not appear anymore. Thanks for a quick response!
The problem was revealed by
http://bugs.freedesktop.org/show_bug.cgi?id=24320
Summary: glXQueryDrawable returns 0 for all attributes except
GLX_FBCONFIG_ID
Product: Mesa
Version: 7.5
Platform: Other
OS/Version: Linux (All)
Status: NEW
http://bugs.freedesktop.org/show_bug.cgi?id=24321
Summary: glXQueryExtension doesn't set eventBase and errorBase
Product: Mesa
Version: 7.5
Platform: Other
OS/Version: Linux (All)
Status: NEW
Severity: normal
http://bugs.freedesktop.org/show_bug.cgi?id=24318
Brian Paul brian.e.p...@gmail.com changed:
What|Removed |Added
Status|NEW |RESOLVED
Chia-I Wu olva...@gmail.com writes:
The first two patches refactor glapi_getproc.c. They provide helper
functions to static dispatches and dynamic dispatches respectively,
and update the rests to use them. There is only one functional
change, the handling of MANGLE. Please see the comments
http://bugs.freedesktop.org/show_bug.cgi?id=24318
--- Comment #3 from Jon TURNEY jon.tur...@dronecode.org.uk 2009-10-05
10:34:51 PST ---
(In reply to comment #2)
NAK!!! The only systems where I have seen M_PI not defined is MSVC. On MSVC
it is not defined /by default/, but it can be
http://bugs.freedesktop.org/show_bug.cgi?id=24318
--- Comment #4 from Brian Paul brian.e.p...@gmail.com 2009-10-05 10:38:27
PST ---
Just a historical note: I recall needing the #ifndef M_PI stuff many years ago
on some Unix variants (HP-UX?), before Mesa was running on Windows.
If you want
Untested, but looks good. This should go on the 7.6 branch.
-Brian
Nicolai Hähnle wrote:
glUniformMatrix() with too large count parameter could previously lead to
memory corruption.
Signed-off-by: Nicolai Hähnle nhaeh...@gmail.com
---
src/mesa/shader/shader_api.c | 29
http://bugs.freedesktop.org/show_bug.cgi?id=24318
--- Comment #5 from Jon TURNEY jon.tur...@dronecode.org.uk 2009-10-05
11:02:33 PST ---
I'll see if I can cook up a patch to do this 'properly' (or at least 'better')
For the compiler/header combination I have, -std=c99 -U__STRICT_ANSI__
http://bugs.freedesktop.org/show_bug.cgi?id=24318
--- Comment #6 from Tom Fogal tfo...@alumni.unh.edu 2009-10-05 11:12:15 PST
---
bugzilla-dae...@freedesktop.org writes:
http://bugs.freedesktop.org/show_bug.cgi?id=24318
--- Comment #2 from Ian Romanick i...@freedesktop.org 2009-10-05
http://bugs.freedesktop.org/show_bug.cgi?id=24334
Summary: mesa git fails to build with xf86vidmodeproto 2.3
Product: Mesa
Version: git
Platform: All
OS/Version: Linux (All)
Status: NEW
Severity: major
Priority:
The texformat-rework branch cleans up the texture format code. The
gl_texture_format struct is replaced by a simple gl_format enum (see
formats.h). This leads to assorted clean-ups in core mesa and the
drivers. This is a step toward unifying and cleaning up the texture
image and renderbuffer
On Mon, Oct 05, 2009 at 07:21:59PM -0700, Eric Anholt wrote:
The third patch fixes a potential segfault. Calling a dynamic dispatch
which is not supported by a DRI driver would crash an application. It
is a bug of the application, since it should test the corresponding
extension before
16 matches
Mail list logo