Lubomir Sedlacik wrote:
> On Tue, Feb 25, 2003 at 01:14:19PM -0500, Nathan J. Williams wrote:
> 
>>The patches are attached; comments requested.
> 
> it's preferred to avoid dependency on auto* tools unless it's really
> necessary.  remove the check from configure instead and integrate it
> into existing patch-ad (or just make the test return needed value, e.g.
> 
> --- configure.orig	2001-03-15 16:44:01.000000000 +0100
> +++ configure	2003-02-25 19:37:01.000000000 +0100

On this Alpha (Feb. 16 userland, Feb. 23 kernel), after applying the above 
method, I still get cores in mozilla, galeon, and evolution--

Galeon:

Program received signal SIGTRAP, Trace/breakpoint trap.
0x16102065c in poll () from /lib/libc.so.12
(gdb) info threads
   3 Thread 0  0x160146604 in _rtld_debug_state () from /usr/libexec/ld.elf_so
* 2 Thread 21  0x16102065c in poll () from /lib/libc.so.12
   1 Thread 22  0x16058ff7c in pthread__locked_switch ()
     from /usr/lib/libpthread.so.0

-----
Mozilla:

Loaded symbols for /steel/X11R6/lib/mozilla/components/libnsprefm.so
#0  0x160322bec in nsGetServiceByCID::operator() ()
from /usr/X11R6/lib/mozilla/libxpcom.so
(gdb) info threads
   8 Thread 0  0x1608de65c in poll () from /lib/libc.so.12
   7 Thread 21  0x1608de65c in poll () from /lib/libc.so.12
   6 Thread 22  0x16088df7c in pthread__locked_switch ()
      from /usr/lib/libpthread.so.0
   5 Thread 23  0x16088df7c in pthread__locked_switch ()
      from /usr/lib/libpthread.so.0
   4 Thread 25  0x16088df7c in pthread__locked_switch ()
      from /usr/lib/libpthread.so.0
   3 Thread 26  0x16088df7c in pthread__locked_switch ()
      from /usr/lib/libpthread.so.0
   2 Thread 27  0x16088df7c in pthread__locked_switch ()
      from /usr/lib/libpthread.so.0
   1 Thread 28  warning: Couldn't find general-purpose registers in core file.

warning: Couldn't find general-purpose registers in core file.

  0x16088df7c in pthread__locked_switch ()
      from /usr/lib/libpthread.so.0

-----
Evolution:

pid 362 (evolution): unaligned access: va=0x1ffffedb4 pc=0x1613f6044 
ra=0x1613f5f24 sp=0x1ffffebf0 op=ldq
(no debugging symbols found)...
Program received signal SIGSEGV, Segmentation fault.
0x161971884 in strlen () from /lib/libc.so.12
(gdb) info threads
* 1 Thread 0  0x161971884 in strlen () from /lib/libc.so.12

Hmm...  Though I *don't* see the gnome-vfs fault I was getting before the glib 
configure patch.  Likely different issues, now.  Should I update userland to 
-current (i.e. are there any recent changes to gdb that can help us chase 
these down?)

Thanks,
Dave