API problems (Was: future versions)

[Waldo Bastian]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu February 19 2004 19:50, Martin Koller wrote:
> On Thursday 19 February 2004 02:01, Michael Brade wrote:
> > Is that the last known KDirLister crash, btw?
>
> Don't know if this is really KDirLister, but have a look at
> http://bugs.kde.org/show_bug.cgi?id=50149

That's pretty much the same issue I guess. The annoying thing is that valgrind 
is not able to catch the problem in a meaningful way, e.g. in one run it 
caught:
==4229== Invalid read of size 4 
 ==4229== at 0x40F78B09: QString::QString(QString const&) (in 
 /usr/lib/qt3/lib/libqt-mt.so.3.1.2) 
 ==4229== by 0x409D1734: KURL::equals(KURL const&, bool) const (in 
 /code/opt/kde3/lib/libkdecore.so.4.2.0) 
 ==4229== by 0x409D16C4: KURL::cmp(KURL const&, bool) const (in 
 /code/opt/kde3/lib/libkdecore.so.4.2.0) 
 ==4229== by 0x405532DE: KDirLister::slotJobToBeKilled(KURL const&) (in 
 /code/opt/kde3/lib/libkio.so.4.2.0) 
<...snip...>
==4229== Address 0x45E9DB94 is not stack'd, malloc'd or free'd 

And in another trace there were already several assert failures before 
valgrind picked up anything :-(

It's also rather hard to reproduce, so I guess it must be a rather specific 
sequence of events, maybe some unexpected recursion from a signal that gets 
emitted??

Cheers,
Waldo
- -- 
bastian at kde.org -=|[ SUSE, The Linux Desktop Experts ]|=- bastian at suse.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFANTRoN4pvrENfboIRAm7mAJ9rFzSpkF7ILhtwIcnjHFWRDXIojACbBk6E
uEtQ63vThKI9k9JzXSk2EnI=
=9Zp4
-----END PGP SIGNATURE-----