A "make clean" combined with removing old stuff no longer referenced by the
makefiles did the trick. Thanks! (I did just "make clean" previously, but that
did not seem to resolve things by itself.)
Morten
Hi there,
on a more-or-less default SuSE 9.3 system I have been getting this
error for a while.
Ideas?
M.
make[3]: Entering directory `/home/welinder/wine/dlls/dinput/tests'
../../../tools/winegcc/winegcc -B../../../tools/winebuild -mconsole joystick.o \
keyboard.o mouse.o testlist.o -o din
Recall that "./wine" usually, but not always, segfaults for me.
I have tracked my race condition down to SHELL_LoadRegistry where the
successful runs exit via the first return ("someone else already loaded the
registry") whereas the segfaulting runs go on.
Ideas?
Morten
From: Michael Stefaniuc <[EMAIL PROTECTED]>
Morten Welinder wrote:
> It looks like it is blowing the stack.
> Ideas?
Let me guess: this is on FC2?
No, as mentioned it's SuSE 9.0.
One more thing: about 1 in 5 runs it turns out that wine does start.
Never und
It looks like it is blowing the stack.
Ideas?
Morten
gdb loader/wine-pthread
...]
(gdb) where
#0 0x4010afdc in _int_malloc () from /lib/libc.so.6
#1 0x4010a18c in malloc () from /lib/libc.so.6
#2 0x404adb72 in _xmalloc (size=25) at ../../misc/registry.c:95
#3 0x404aeb24 in _w95_dump_dke
No change with 'set follow-fork-mode child', or even '... ask'.
With strace, I get...
strace -i $WINELOADER
[...]
[40113d71] close(9) = 0
[400f0c37] getpid() = 30133
[40113ce3] open("/tmp/reg75b5.tmp", O_WRONLY|O_CREAT|O_EXCL|O_LARGEFILE, 0666) = 9
[40
A newly built wine segfaults for me. No args:
[EMAIL PROTECTED]:~/wine> ./wine
Segmentation fault
No core. gdb just reports:
[EMAIL PROTECTED]:~/wine> LD_LIBRARY_PATH=/home/welinder/wine/libs/wine gdb
loader/wine
...
(gdb) r
Starting program: /home/welinder/wine/loader/wine
[New Thread 16