|
From: Julian S. <js...@ac...> - 2006-06-24 12:28:43
|
> ==5606== Conditional jump or move depends on uninitialised value(s) > ==5606== at 0x5486770: KDirListerCache::DirItem::incAutoUpdate() > (kdirlister_p.h:285) ==5606== by 0x548543B: > KDirListerCache::listDir(KDirLister*, KURL const&, bool, bool) > (kdirlister.cpp:271) ==5606== by 0x54855A2: KDirLister::openURL(KURL > const&, bool, bool) (kdirlister.cpp:1863) ==5606== by 0xD3885FF: > KonqBaseListViewWidget::openURL(KURL const&) (konq_listviewwidget.cc:1025) > > which all looks like false positives. I know that you suggest to use only > -O instead of -O2 but e.g. the first issue is from X.Org which is from > Debian and I don't want to compile it myself. Is there another workaround > or could you add a new option to lets work memcheck more accurate (even if > it's slower)? Try using 3.2.0; I put a bunch of fixes in there to reduce the false error rate. J |
|
From: <Woe...@on...> - 2006-06-24 12:39:17
|
On Saturday 24 June 2006 14:28, Julian Seward wrote: > > =3D=3D5606=3D=3D Conditional jump or move depends on uninitialised valu= e(s) > > =3D=3D5606=3D=3D at 0x5486770: KDirListerCache::DirItem::incAutoUpda= te() > > (kdirlister_p.h:285) =3D=3D5606=3D=3D by 0x548543B: > > KDirListerCache::listDir(KDirLister*, KURL const&, bool, bool) > > (kdirlister.cpp:271) =3D=3D5606=3D=3D by 0x54855A2: > > KDirLister::openURL(KURL const&, bool, bool) (kdirlister.cpp:1863) > > =3D=3D5606=3D=3D by 0xD3885FF: KonqBaseListViewWidget::openURL(KURL > > const&) (konq_listviewwidget.cc:1025) > > > > which all looks like false positives. I know that you suggest to > > use only -O instead of -O2 but e.g. the first issue is from X.Org > > which is from Debian and I don't want to compile it myself. Is > > there another workaround or could you add a new option to lets work > > memcheck more accurate (even if it's slower)? > > Try using 3.2.0; I put a bunch of fixes in there to reduce the > false error rate. yeah, I read this in the changelog and tried it as soon as there was a=20 Debian package in the hope for better results. But unfortunately the=20 false positives were still there and I wrote the mail :-( Sorry for the lack of information. Cheers, Andr=E9 |