Ville Syrjälä wrote:
On Sat, Nov 26, 2005 at 08:24:42PM +0000, James Gatt wrote:
Ville Syrjälä wrote:
Is there any debugging output I can enable from DirectFB (especially the
matrox driver) which will help?
You could compile with --enable-trace which should give a backtrace
without gdb.
AFAICS the only place in maven_init() where it could segfault is
direct_memcpy(). But I don't think it's actually segfaulting there
because then the segfault should happen when calling DirectFBCreate().
I think maven_init() might be a red herring... I've rebuilt DirectFB
from source (not using a Gentoo ebuild) and it's crashing when calling
RGB_TO_YCBCR() in the second iteration of the loop in spicSetRegion().
RGB_TO_YCBCR() was changed to use pre-computed tables. The tables
are indexed using values the can be negative. Maybe that is the cause
for the segfault.
Try this patch:
Index: src/gfx/convert.h
Index: src/gfx/generic/yuvtbl-gen.c
Index: src/gfx/generic/yuvtbl.h
Thanks, that's fixed it. :)
I'd really like to get those fixes into the Gentoo ebuild. Assuming
there isn't going to be another DirectFB release for a while, is there
someone in the DirectFB project who normally deals with Gentoo?
If there isn't I could always send them the diffs for the three files
above and ask them to make a patch, but I don't want to tread on
anyone's feet.
Regards,
James.
_______________________________________________
directfb-users mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
