[isf-wifidog] output

[acv]

On Tue, Jul 29, 2008 at 01:41:04PM -0400, Clifford Thurber wrote:
> Date: Tue, 29 Jul 2008 13:41:04 -0400
> From: "Clifford Thurber" <clifford at hdn.net>
> To: acv at miniguru.ca
> Subject: output
>
> Hi Alex,
> This is what I am seeing now:
> 
[ ... snip ... ]
> called. Trying to reap a child
> [7][Mon Jul 28 14:47:27 2008][26253](gateway.c:263) Handler for SIGCHLD
> reaped child PID -1
> [7][Mon Jul 28 14:47:27 2008][26253](util.c:121) Process PID 26358 exited
> [7][Mon Jul 28 14:47:27 2008][26253](centralserver.c:327) Level 1:
> Connecting to auth server auth.ilesansfil.org:80
> *** glibc detected *** wifidog: free(): invalid pointer: 0x0000003ea656a0c0
> ***
> ======= Backtrace: =========
> /lib64/libc.so.6[0x3ea6278158]
> /lib64/libc.so.6(cfree+0x76)[0x3ea627a796]
> wifidog[0x4081ef]
> wifidog[0x407d35]
> wifidog[0x40baab]

Congratulations, you found a hidden bug. Most versions of libc silently
ignore when an application tries to free() memory that's already been
free()'d. Gentoo, Ubuntu and apparently now Fedora have started enabling
optional behaviour in the free() system call to exit out when that happens.
This was originally implemented for debugging purpose.

Now we have to find where free() gets called twice... Probably when connecting
to the Auth Server.

Alex
-------------- section suivante --------------
Une pièce jointe non texte a été nettoyée...
Nom: non disponible
Type: application/pgp-signature
Taille: 187 octets
Desc: non disponible
Url: http://listes.ilesansfil.org/pipermail/wifidog/attachments/20080729/54eaec93/attachment.pgp