[no subject]

[snip]

You second mail celars up some things. The MMX code isn't the problem it
seems (since you compiled it out). Secdonly, the backtrace reveals that
it's the call to clear(screen) inside set_gfx_mode() that is causeing
the problem.

I'm not expert in X though, so I can't really say what's going on.

[FW: second mail]
Some more details : by disabling totally mmx in the library, I get a
new backtrace (of course) :

Program received signal SIGSEGV, Segmentation fault.
0x400a448f in _xwin_write_line (bmp=0x4e08002b, line=-1048836090)
    at ./src/x/xwin.c:2548
2548       int new_line = line + bmp->y_ofs;
(gdb) bt
#0  0x400a448f in _xwin_write_line (bmp=0x4e08002b, line=-1048836090)
    at ./src/x/xwin.c:2548
#1  0x804c2cd in clear_loop ()
#2  0x40099cf3 in _xwin_clear_to_color (dst=0x8064e08, color=0)
    at ./src/x/xvtable.c:251
#3  0x4005af73 in clear (bitmap=0x8064e08) at
include/allegro/alinline.h:699
#4  0x4005f552 in set_gfx_mode (card=0, w=320, h=200, v_w=0, v_h=0)
    at ./src/graphics.c:563
#5  0x4005ecc7 in set_gfx_mode (card=1396786757, w=320, h=200, v_w=0,
v_h=0)
    at ./src/graphics.c:341
#6  0x804c165 in _mangled_main () at ./examples/exmem.c:23

clear_loop is in iblit8.s. I don't feel like debuging assembly right
now, so now I am left with trying to remove all the assembly.. I think
it should work, and be faster than the debug library...



--
- Robert J Ohannessian