Haven't seen this one yet, but this is very critical if it happens for more users. Could you try the following?

1) delete $HOME/.kde/share/apps/akregator/Archive/feedlistbackup.mk4

2) Check some of the backtraces whether "storeFeedList" shows up in all of them

Created attachment 13171 [details]
Backtraces

I deleted $HOME/.kde/share/apps/akregator/Archive/feedlistbackup.mk4 but that didn't change anything.

Here are some more backtraces, "storeFeedList" is always in there.
If I can somehow get more information for you please let me know.
Also this happens without me doing anything, it can just be minimized.

*** Bug 115246 has been marked as a duplicate of this bug. ***

Interestingly both reporters use Gentoo. Thomas, could you also provide your CFLAGS setting?

These are from #115246:

CFLAGS="-O2 -mcpu=i686 -march=pentium3 -mfpmath=sse -msse2 -mmmx" 

I can make it crashes always closing it from the system tray.For you it's the same?

It crashes only if you close it from systray but not when you choose File->Quit from the menu?? The actions performed are exactly the same.

Yup, it crashes closing it (all the way). I found a user in GENTOO who compiled it with GCC 4, and got no problem...

I got the same crashes, also using Gentoo. I don't know hwo to reproduce it, I don't even have the trayicon enabled. Closing it from the Akregator menu doesn't produce a crash. It just crashed twice today. Once inside Kontact, no crashreport, Kontact just disappeared. Second time in Kontact I did get a backtrace. I have been running 3.5 beta2 since it came available.

KDE Version: 1.2 (KDE 3.4.92 (beta2, >= 20051010), Gentoo)
Application: KDE feed aggregator
Operating: System  Linux (i686) release 2.6.13-suspend2-r5
Compiler: gcc version 3.3.6 (Gentoo 3.3.6, ssp-3.3.6-1.0, pie-8.7.8)

CFLAGS="-Os -march=athlon-xp -pipe -ggdb3"

Backtrace:

Using host libthread_db library "/lib/libthread_db.so.1".
`system-supplied DSO at 0xffffe000' has disappeared; keeping its symbols.
[Thread debugging using libthread_db enabled]
[New Thread -1238619424 (LWP 20746)]
[KCrash handler]
#4  c4_FormatB::ItemLenOffCol (this=0x8a7b658, index_=0, off_=@0xbfc25f3c, 
    col_=@0xbfc25f40) at univ.inl:41
#5  0xb582904c in c4_FormatB::GetOne (this=0x8a7b658, index_=0, 
    length_=@0xbfc25f84) at format.cpp:623
#6  0xb58298eb in c4_FormatS::Get (this=0x8a7b658, index_=0, 
    length_=@0xbfc25f84) at format.cpp:896
#7  0xb582a7ea in c4_Handler::GetBytes (this=0xbfc25f40, index_=0, 
    buf_=@0xbfc2601c, copySmall_=false) at handler.cpp:58
#8  0xb5836517 in c4_Sequence::Get (this=0x8a7b150, index_=0, propId_=1, 
    buf_=@0xbfc2601c) at viewx.cpp:356
#9  0xb582ebd0 in c4_HashViewer::CalcHash (this=0x8a7b978, cursor_=
      {_seq = 0x8a7b150, _index = 0}) at handler.inl:35
#10 0xb582ef11 in c4_HashViewer::InsertDict (this=0x8a7b978, row_=0)
    at remap.cpp:302
#11 0xb582f3ff in c4_HashViewer::SetItem (this=0x8a7b978, row_=0, col_=0, 
    buf_=@0xbfc261b0) at remap.cpp:423
#12 0xb5822240 in c4_CustomSeq::DoSet (this=0x0, row_=0, col_=0, 
    buf_=@0xbfc261b0) at custom.cpp:178
#13 0xb5821f05 in c4_CustomHandler::Set (this=0x8a6c860, index_=0, 
    buf_=@0xbfc261b0) at custom.cpp:77
#14 0xb5835cf2 in c4_Sequence::SetAt (this=0x8a79c20, index_=0, newElem_=
      {_seq = 0x820cd98, _index = 0}) at viewx.cpp:116
#15 0xb581c3ca in Akregator::Backend::StorageMK4Impl::storeFeedList (
    this=0x8a7a318, opmlStr=@0xbfc2630c) at mk4.inl:13
#16 0xb558adc4 in Akregator::Part::slotSaveFeedList (this=0x8a64508)
    at akregator_part.cpp:443
#17 0xb558d450 in Akregator::Part::qt_invoke (this=0x8a64508, _id=6, 
    _o=0xbfc263c8) at akregator_part.moc:212
#18 0xb6a13053 in QObject::activate_signal (this=0x8ae3580, clist=0x8a7be70, 
    o=0xbfc263c8) at qconnection.h:56
#19 0xb6a12e94 in QObject::activate_signal (this=0x8a7b658, 
    signal=-1077779512) at qobject.cpp:2324
#20 0xb6d490ab in QTimer::timeout (this=0x0) at qmetaobject.h:261
#21 0xb6a33fe0 in QTimer::event (this=0x8ae3580, e=0x0) at qtimer.cpp:219
#22 0xb69b7b3c in QApplication::internalNotify (this=0x0, receiver=0x8ae3580, 
    e=0xbfc26810) at qapplication.cpp:2615
#23 0xb69b6edd in QApplication::notify (this=0xbfc26a44, receiver=0x8ae3580, 
    e=0xbfc26810) at qapplication.cpp:2372
#24 0xb7018ea4 in KApplication::notify (this=0xbfc26a44, receiver=0x8ae3580, 
    event=0xbfc26810) at kapplication.cpp:550
#25 0xb69a75c9 in QEventLoop::activateTimers (this=0x8165478)
    at qeventloop_unix.cpp:555
#26 0xb69618c7 in QEventLoop::processEvents (this=0x8165478, flags=4)
    at qeventloop_x11.cpp:389
#27 0xb69c9a98 in QEventLoop::enterLoop (this=0x8165478) at qeventloop.cpp:198
#28 0xb69c9948 in QEventLoop::exec (this=0x8165478) at qeventloop.cpp:145
#29 0xb69b7d71 in QApplication::exec (this=0xbfc26a44)
    at qapplication.cpp:2758
#30 0x080592f5 in main (argc=1, argv=0xbfc26c04) at main.cpp:177

Sorry for my late reply I had a nasty hd crash.

Here are my cflags: CFLAGS="-march=athlon-xp -O2"

Here my gcc and kernel version: gcc-4.0.2, glibc-2.3.5.20050725-r0, 2.6.13-dark2 i686

I don't have the crash bug on closing it though.
I can right click on the icon and quit it without problems.

My crashes are totally random though it seems to have the same cause.

Any clue what might cause this? I have seen a couple of crashes with this backtrace already. Crashes ~twice per session.

I don't see anything common in CFLAGS, gcc version etc. All reporters use Gentoo though. Did anyone of you use 3.5 Beta1 or an SVN checkout before? If yes, with or without crashes?

Yes, I have used 3.5 beta1, without crashes.

SVN commit 478644 by osterfeld:

trying to fix the problems with storeFeedList()
Gentoo users, it would be great if someone could apply this patch and check if it solves the problem.
CCBUG: 114831
CCMAIL: zander@kde.org


 M  +24 -12    storagemk4impl.cpp  


--- branches/KDE/3.5/kdepim/akregator/src/mk4storage/storagemk4impl.cpp #478643:478644
@@ -316,14 +316,20 @@
 
 void StorageMK4Impl::storeFeedList(const QString& opmlStr)
 {
-    c4_Row row;
-
+   
     if  (d->feedListView.GetSize() == 0)
+    {
+        c4_Row row;
+        d->pFeedList(row) = !opmlStr.isEmpty() ? opmlStr.utf8().data() : "";
+        d->pTagSet(row) = "";
         d->feedListView.Add(row);
-    
-    row = d->feedListView.GetAt(0);
-    d->pFeedList(row) = opmlStr.utf8().data();
-    d->feedListView.SetAt(0, row);
+    }
+    else
+    {
+        c4_Row row = d->feedListView.GetAt(0);
+        d->pFeedList(row) = !opmlStr.isEmpty() ? opmlStr.utf8().data() : "";
+        d->feedListView.SetAt(0, row);
+    }
     d->modified = true;
 }
 
@@ -338,14 +344,20 @@
 
 void StorageMK4Impl::storeTagSet(const QString& xmlStr)
 {
-    c4_Row row;
-
+   
     if  (d->feedListView.GetSize() == 0)
+    {
+        c4_Row row;
+        d->pTagSet(row) = !xmlStr.isEmpty() ? xmlStr.utf8().data() : "";
+        d->pFeedList(row) = "";
         d->feedListView.Add(row);
-
-    row = d->feedListView.GetAt(0);
-    d->pTagSet(row) = xmlStr.utf8().data();
-    d->feedListView.SetAt(0, row);
+    }
+    else
+    {
+        c4_Row row = d->feedListView.GetAt(0);
+        d->pTagSet(row) = !xmlStr.isEmpty() ? xmlStr.utf8().data() : "";
+        d->feedListView.SetAt(0, row);
+    }
     d->modified = true;
 }
 

I compiled it and use it now.
So far no crash so I hope it's fixed.
But I will let it run for a few days and report back.

I am using Gentoo.

SVN commit 478672 by osterfeld:

this line mixes feed archive and feed list backup and does not make sense at all. Let's see if removing it
fixes 114831. Actually I wonder why this code worked at all for at least some people...
CCBUG: 114831
CCMAIL: zander@kde.org


 M  +0 -1      storagemk4impl.cpp  


--- branches/KDE/3.5/kdepim/akregator/src/mk4storage/storagemk4impl.cpp #478671:478672
@@ -111,7 +111,6 @@
     filePath = d->archivePath +"/feedlistbackup.mk4";
     d->feedListStorage = new c4_Storage(filePath.local8Bit(), true);
     d->feedListView = d->feedListStorage->GetAs("archive[feedList:S,tagSet:S]");
-    d->feedListView = d->feedListView.Hash(hash, 1); // hash on url
     return true;
 }
 

Ok, I didn't have any crashes anymore and akregator was running the whole day.

I am pretty sure it's fixed.

Thanks a lot.

Did you apply the patch from #14 only or also the patch from #16?

I just checked it out from svn again.
First with the #14 patch and then after I saw you submitted another patch I checked it out from svn again and recompiled it.

Seems fixed, let's close it