did arena allocations gain the speed ?

Alfons Hoogervorst alfons at proteus.demon.nl
Sat Mar 6 16:23:03 GMT 2004


Lo Dirk,

On 06-03-04 (Sat) 15:39 +0100 Dirk Mueller wrote: 

| > (latest CVS, kdelibs/khtml/khtml_part.cpp +402):
| 
| and it appeared on HEAD 2 days ago? How to reproduce?

Make sure you have cleaned your cache; then go to http://newsforge.com,
there's a blurb about Computer Associates and Linux. Click on the link
called "Read more at Davanum Srinivas' weblog". One out of the 3 times
this gets me a nice bt:

#1  0x41cb2384 in ~KHTMLPart (this=0x86a58c8) at khtml_part.cpp:402
#2  0x40270aaa in KParts::Part::slotWidgetDestroyed() (this=0x86a58c8)
at part.cpp:267
#3  0x40272895 in KParts::Part::qt_invoke(int, QUObject*)
(this=0x86a58c8, _id=2, _o=0xbfffdc50) at part.moc:108
#4  0x40272cdb in KParts::ReadOnlyPart::qt_invoke(int, QUObject*)
(this=0x86a58c8, _id=2, _o=0xbfffdc50) at part.moc:261
#5  0x41cce642 in KHTMLPart::qt_invoke(int, QUObject*) (this=0x86a58c8,
_id=2, _o=0xbfffdc50) at khtml_part.moc:531
#6  0x40f0fa5e in QObject::activate_signal(QConnectionList*, QUObject*)
() from /opt3/qt-copy/lib/libqt-mt.so.3

I've looked at an hour or so at this and it really seems to me that the
pointer to the KHTMLPartPrivate (d) is invalid, sometimes just zero (the
destructor zeros the pointer).

Bye.

-- 
Ecuación algebraico sin solución posible,
a menos de poseer profundos conocimientos
en matemática - Revueltas (Ocho Por Radio)




More information about the kfm-devel mailing list