From: Allan Schaffer (allan++at++sgi.com)
Date: 03/18/2002 20:06:30
Hello Ive,
Ive Akkermans wrote:
> PF Warning/Internal(12): gluBuild2DMipmaps failed on red.sg
> PF Defaulting to bilinear.
>
> I don't get this error all the time, and not always with this texture.
> When I use gdb for instance, it produces the same error but on 2 other
> textures.
>
> In any case, when this error is produced, my app segfaults too:
>
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread 1024 (LWP 20168)]
> 0x40c1cb00 in gluScaleImage () from /usr/lib/libGLU.so.1
> Current language: auto; currently c
> (gdb) info stack
> #0 0x40c1cb00 in gluScaleImage () from /usr/lib/libGLU.so.1
> #1 0x00000003 in ?? ()
> Cannot access memory at address 0xc1b7d9ac
This sort of error (where the stack starts at nothing) usually comes
from a NULL C++ object member being dereferenced, e.g. conceptually:
ClassType *foo = NULL;
foo->setBar(bar);
Check around your code for cases where a return value might not be
getting checked.
In Performer, gluBuild2DMipmaps gets called from pfTexture::pr_format,
which is an internal routine that preps a texture. I don't see a call
to gluScaleImage() anywhere though.
My offhand guess is that some texture is coming back NULL but is being
passed along later to be used.
Octane's & Onyx's allow programs to dereference NULL without a
segmentation violation so maybe you're just getting lucky.
Good luck. :-)
Allan
-- Allan Schaffer allan++at++sgi.com Core Rendering 1-650-933-2160 Silicon Graphics http://www.sgi.com
This archive was generated by hypermail 2b29 : Mon Mar 18 2002 - 20:07:30 PST