[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On Fri, 24 Nov 2000, Angelo Mottola wrote:
> There is something wrong with the latest CVS changes... Current 3.9.34 WIP
> crashes under Linux (tried it under XF4) with a segmentation fault. In case
> you want to reproduce the sigsegv, here's the way I did it:
> Open the grabber and load demo.dat; now delete some random items on the
> objects list, and the program will crash...
I've checked it and for me it crashes on another point. If I run
"./grabber -16 ../demo/demo.dat", grabber will first set a 16 bpp mode,
then load the datafile. At the same moment I try to show the first item
(an rle sprite), it crashes. I compiled against the debug version, and the
culprit is src/color.c:
Program terminated with signal 11, Violación de segmento.
#0 0x8066b24 in create_rgb_table (table=0x540, pal=0x555, callback=0xa800) at ./src/color.c:557
557 dopos(-1, 0, 0, 1);
(gdb) bt
#0 0x8066b24 in create_rgb_table (table=0x540, pal=0x555, callback=0xa800) at ./src/color.c:557
#1 0x15 in ?? ()
Cannot access memory at address 0xf81f.
The dopos macro is too convulted and obscure for me to understand what
it's all about, but it looks like a simple buffer overrun (either reading
or writting). Grabber also crashes with 32 bpp, but not 8 or 24. Why?
Could somebody check this? I presume gdb can't access frame #1 because
plugins don't seem to be compiled with debug info, and it's the only thing
IIRC I left without compiling manually.
> This has never happened before with 3.9.33, so it should be due to some wierd
> new changes.
I checked the cvs-web interface, and there weren't significative changes
since 3.9.33, so maybe it's just now that it happens to crash, under a
determinate environment.
Grzegorz Adam Hankiewicz gradha@xxxxxxxxxx - http://gradha.infierno.org
Other web pages: http://glub.ehu.es/ - http://welcome.to/gogosoftware/