Mail Archives: cygwin-developers/2002/01/16/23:48:59
----- Original Message -----
From: "Christopher Faylor" <cgf AT redhat DOT com>
> On Thu, Jan 17, 2002 at 03:18:52PM +1100, Robert Collins wrote:
> >----- Original Message -----
> >From: "Christopher Faylor" <cgf AT redhat DOT com>
> You have to set some break points. You're single stepping in an
> location in an unknown thread.
I'm aware of that :}. Thats why I tried stepi - which should work
> You probably also have to switch to the correct thread. It may be
> thread 1 but with ttys there are more threads involved.
There are three threads active - all in system
thread 1688.0x694 0x77f7f571 in _system_dlls__ ()
thread 1688.0xc38 0x7ffe0304 in ?? ()
thread 1688.0xccc 0x7ffe0304 in ?? ()
> info threads will display the thread. "thread 1" will switch to
> it. "bt" will show a back trace. It's possible that the back trace
> will be inaccurate in some cases due to system dll "frame pointerless"
Yep. thread 1 and 3 has that problem, the other one shows
#0 0x7ffe0304 in ?? ()
#1 0x77e74c70 in _system_dlls__ ()
#2 0x61006285 in thread_stub (arg=0x610a6930)
> >Program exited with code 0200400.
> > ^ ^^^^^^^^^^^^
> >How should this code be interpreted?
> dcrt0.cc: do_exit()
Ok that helps a bit, it's exiting with REPARENTING set (which is very
weird as sh should be vforking, not using _P_OVERLAY), but I still don't
grok 0x400 as a exit value. it seems a little large:}.
> >(gdb) attach 2652
> >Attaching to program `/bin/sh.exe', process 2652
> >[Switching to thread 2652.0xbb0]
> >(gdb) stepi
> >0x77f970a4 in _system_dlls__ ()
> 0x77* is a system DLL.
> I think the command is "info shared" to list DLLs but since my cygwin
> installation is currently hosed, I can't check.
It's in ntdll.dll, the kernel API gateway.
- Raw text -