Re: crash when editing from a search in certain conditions

[wwp <subscript free fr>]

Hello Pavel,


On Thu, 7 Sep 2006 10:22:33 +0300 (EEST) Pavel Tsekov <ptsekov gmx net> wrote:

> On Wed, 6 Sep 2006, wwp wrote:
>
> > using mc 2006-08-12-18 on a FC5 box, I get a crash w/ a specific scenario:
> >
> > - be in /tmp/anydiryoulike
> > - create a subdir es/ (/tmp/anydiryoulike/es/)
> > - touch es/foobar
> > - launch mc
> > - search for file * (default search options)
> > - search results should show: es, foobar in es/
> > - move the cursor on es/foobar, press F4 to edit
> > - bang
> >
> > It seems that the subdir name (es/) matters!
>
> Using MC from cvs I cannot reproduce the crash. Here is
> the output of "mc -V":
>
> GNU Midnight Commander 4.6.1a
> Virtual File System: tarfs, extfs, cpiofs, ftpfs, fish
> With builtin Editor
> Using included S-Lang library with terminfo database
> With optional subshell support
> With support for background operations
> With mouse support on xterm and Linux console
> With support for X11 events
> Data types: char 8 int 32 long 32 void * 32 off_t 64 ecs_char 8
>
> > Is there anything I can provide more than this incomplete backtrace (mc
> > compiled w/ -g, `make install` no strip).
>
> CFLAGS="-g -O0" ./configure ; make
>
> > Program received signal SIGABRT, Aborted.
> > 0x0060a402 in __kernel_vsyscall ()
> > (gdb)
> > (gdb) bt full
> > #0  0x0060a402 in __kernel_vsyscall ()
> > No symbol table info available.
> > #1  0x00650069 in raise () from /lib/libc.so.6
> > No symbol table info available.
> > #2  0x00651671 in abort () from /lib/libc.so.6
> > No symbol table info available.
> > #3  0x00684a4b in __libc_message () from /lib/libc.so.6
> > No symbol table info available.
> > #4  0x0068c660 in _int_malloc () from /lib/libc.so.6
> > No symbol table info available.
> > #5  0x0068dc74 in malloc () from /lib/libc.so.6
>
> Seems like a call to malloc() fails. Maybe some chunk of
> memory was overwritten or misused in some other way, maybe
> it is just lack of memory. Does it crash consistently ?

Bug can't be reproduced:
 - with FC5's mc (IIRC, 20060812-18 snapshot + FC patches)
 - with 20060307-16 snapshot (-O2)
 - current CVS (-O2)

bug can be reproduced (100% w/ the scenario above):
 - with 20060812-18 snapshot, with -O0 or -O2

I think I'll keep the CVS one ;-).


Regards,

-- 
wwp

Attachment: signature.asc
Description: PGP signature