This is the mail archive of the
cygwin@cygwin.com
mailing list for the Cygwin project.
Re: Select error and stackdumps
On Wed, Jun 06, 2001 at 10:21:05AM -0500, Joey Mukherjee wrote:
>
> >At 06:34 PM 6/5/2001, Joey Mukherjee wrote:
> >>What can I do with a stack dump? Its not like a core since I tried loading it
> >>into gdb on the command line, but maybe there is another more obvious way to
> >>look at it that I'm missing.
> >
> >Yes, look at it with your eyes! ;-) Its ASCII.
>
> I did, but how you do make sense of the numbers:
>
> Exception: STATUS_ACCESS_VIOLATION at eip=0045C115
> eax=00000000 ebx=00000004 ecx=610902DC edx=61095320 esi=610903E8 edi=00000008
> ebp=026CF454 esp=026CF434 program=f:\CYGWIN\sddas\bin\SpectroScalar.exe
> cs=001B ds=0023 es=0023 fs=003B gs=0000 ss=0023
> Stack trace:
> Frame Function Args
> 026CF454 0045C115 (0A03EE30, 0A064F90, 000003D3, 00000000)
> 026CFEB4 00401A0F (00000001, 1A026374, 0A010008, 00000000)
> 026CFF10 61003AEA (00000000, 0256CFB4, 0715FBB4, 00000004)
> 026CFF40 61003CBD (0040131C, 0256CFB4, 847B75C0, 804A2D00)
> 026CFF60 61003CFC (00000000, 00000000, 847B7750, 00000005)
> 026CFF90 00648C73 (0040131C, FFFFFFFF, 80430C77, 00000000)
> 026CFFC0 0040103D (0256CFB4, 0256D234, 7FFDF000, 1A02397C)
> 026CFFF0 77E992A6 (00401000, 00000000, 000000C8, 00000100)
>
> How can I convert this to something which would work like a core file? Is it
> possible? I already knew which program was bombing out. Can I trace this back
> to a statement in my program?
You can use gdb. If your program is compiled with debug symbols (-g)
and not stripped you can at least find the functions since gdb
knows of course the function-address relation. The above addresses
show that the error happens in your own code. Function addresses
in the 0x61000000 address space are Cygwin functions, the 0x77E...
is a OS function.
Corinna
--
Corinna Vinschen Please, send mails regarding Cygwin to
Cygwin Developer mailto:cygwin@cygwin.com
Red Hat, Inc.
--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple