Segfault under cygwin 1.7.62

[Vin Shelton]

Stephen -

Thanks again for taking this up.

On Thu, Oct 8, 2009 at 2:36 PM, Stephen J. Turnbull <stephen@xemacs.org> wrote:
> Vin Shelton writes:
>
>  > (gdb) pobj charset
>  > Cannot access memory at address 0x4
>
> Urk.  That's unexpected.  You might want to try going up the stack to
> frame 1 and trying that.  If that's not more sensible, what does print
> charset say?

(gdb) frame 1
#1  0x005f05de in unicode_convert (str=0x1710a48,
    src=0x148d123
"ï\202\201Ð\201Ð\201Ð\201ï\203\2204VB3Upi-string)g)", dst=0x14e8b40,
    n=17) at /usr/local/src/xemacs-21.5-test-2009-10-07-mule/src/charset.h:247
247     DECLARE_LRECORD (charset, Lisp_Charset);
(gdb) pobj charset
$5 = (struct Lisp_Symbol *) 0x665cb0
$6 = {lheader = {type = 4, mark = 0, c_readonly = 0, lisp_readonly =
0, uid = 260},
  next = 0x0, name = 7151736, value = 6708400, function = 6401080,
plist = 6708400}
Symbol name: nil

>
> Diving into the stack
>
>  > #9  0x0056d4d7 in qxe_realpath (path=0x22248c "",
>  >     resolved_path=0x22651c
>  > "c:/cygwin/tmp/s360339.aoa/\201Ð\201Ð\201Ð\201Ð02ABFx",
>  >     links_only=0) at
>  > /usr/local/src/xemacs-21.5-test-2009-10-07-mule/src/realpath.c:348
>
> Does the filename above make sense to you?  Is there a
> "c:/cygwin/tmp/s360339.aoa" directory?  Does it contain non-ASCII
> filenames?

Yes and yes.  That would be my user-specific tmp directory.

>
>
>  > #8  0x0056cd68 in readlink_or_correct_case (name=<value optimized out>,
>  >     buf=0x21e464
>  > "/tmp/s360339.aoa/\201ï\217£\217°\201ï\217£\217°\201ï\217£\217°\201ï\217£\217°02ABFx",
>  > size=16383, links_only=0)
>  >     at /usr/local/src/xemacs-21.5-test-2009-10-07-mule/src/realpath.c:176
>
> Urk the representation of the filename has changed ...
>
>  > #3  0x005151bc in Lstream_really_write (lstr=0x176ea00,
>  >     data=0x148d0c0
>  > "c;C:\\cygwin\\cygwin\\tmp\\s360339.aoa\\ï\202\201Ð\201Ð\201Ð\201ï\203\22002ABFxi-string)g)Àâw\001\200ÑH\001Â\200\t",
>  > size=53)
>  >     at /usr/local/src/xemacs-21.5-test-2009-10-07-mule/src/lstream.c:380
>
> and it changes back to something similar but not identical to the
> original.  Note the "ï\202" that got inserted somehow.  The original
> stuff at frame #9 is Mule internal coding, but this is not.

Ahh!  Corinna - are these changes in filename a result of your changes
in 1.7.62?  Any thoughts?

Regards,
  Vin

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple