you better run it in valgrind and stop reporting a duplicate each day :)

Subject: Re:  another crash in kacctimap         

On 25 Nov 2003, Stephan Kulow wrote:

> you better run it in valgrind and stop reporting a duplicate each day :)
Can do, i am running kde head for 4 users on the same machine. They are 
really good testers :-)

This is the result of the polish jury: Germany 2 points ;)

==7545== Invalid read of size 4
==7545==    at 0x58F3FB: QValueListIterator<QGuardedPtr<KMFolder> >::operator++(int) (qvaluelist.h:121)
==7545==    by 0x69B73C: KMAcctImap::processNewMail(bool) (kmacctimap.cpp:300)
==7545==    by 0x64A784: KMAcctMgr::processNextCheck(bool) (kmacctmgr.cpp:184)
==7545==    by 0x64A115: KMAcctMgr::singleCheckMail(KMAccount*, bool) (kmacctmgr.cpp:118)
==7545==    by 0x64ABE1: KMAcctMgr::checkMail(bool) (kmacctmgr.cpp:286)
==7545==    by 0x733FD6: KMMainWidget::slotCheckMail() (kmmainwidget.cpp:735)
==7545==    by 0x731E3C: KMMainWidget::readConfig() (kmmainwidget.cpp:331)
==7545==    by 0x730B6E: KMMainWidget::KMMainWidget(QWidget*, char const*, KActionCollection*, KConfig*) (kmmainwidget.c
pp:118)
==7545==    by 0x5686C7: KMMainWin::KMMainWin(QWidget*) (kmmainwin.cpp:19)
==7545==    by 0x6CE8A8: KMKernel::openReader() (kmkernel.cpp:198)
==7545==    by 0x6D35FF: KMKernel::action(bool, bool, QString const&, QString const&, QString const&, QString const&, QS
tring const&, KURL const&, KURL::List const&) (kmkernel.cpp:1262)
==7545==    by 0x804C5BD: KMailApplication::newInstance() (main.cpp:159)
==7545==    by 0x678DF65: KUniqueApplication::newInstanceNoFork() (kuniqueapplication.cpp:343)
==7545==    by 0x678E5C8: KUniqueApplication::qt_invoke(int, QUObject*) (kuniqueapplication.moc:87)
==7545==    Address 0x42455838 is 0 bytes inside a block of size 12 free'd
==7545==    at 0xA41DFD: __builtin_delete (vg_replace_malloc.c:244)
==7545==    by 0xA41E28: operator delete(void*) (vg_replace_malloc.c:253)
==7545==    by 0x5AAA87: QValueListPrivate<QGuardedPtr<KMFolder> >::clear() (qvaluelist.h:386)
==7545==    by 0x5A90BC: QValueList<QGuardedPtr<KMFolder> >::clear() (qvaluelist.h:590)
==7545==    by 0x69B82B: KMAcctImap::slotUpdateFolderList() (kmacctimap.cpp:346)
==7545==    by 0x69BD1A: KMAcctImap::qt_invoke(int, QUObject*) (kmacctimap.moc:113)
==7545==    by 0x17441A5: QObject::activate_signal(QConnectionList*, QUObject*) (in /usr/local/qt-copy/lib/libqt-mt.so.3
.2.2)
==7545==    by 0x1744063: QObject::activate_signal(int) (in /usr/local/qt-copy/lib/libqt-mt.so.3.2.2)
==7545==    by 0x6558F1: KMFolderMgr::changed() (kmfoldermgr.moc:130)
==7545==    by 0x654E20: KMFolderMgr::contentsChanged() (kmfoldermgr.cpp:347)
==7545==    by 0x654B23: KMFolderMgr::remove(KMFolder*) (kmfoldermgr.cpp:302)
==7545==    by 0x6B3709: KMFolderImap::processNewMail(bool) (kmfolderimap.cpp:1384)
==7545==    by 0x69B705: KMAcctImap::processNewMail(bool) (kmacctimap.cpp:318)
==7545==    by 0x64A784: KMAcctMgr::processNextCheck(bool) (kmacctmgr.cpp:184)
==7545==


==7545== Invalid read of size 4
==7545==    at 0x58F7D4: QGuardedPtrPrivate::object() const (qguardedptr.h:136)
==7545==    by 0x69B62B: KMAcctImap::processNewMail(bool) (kmacctimap.cpp:302)
==7545==    by 0x64A784: KMAcctMgr::processNextCheck(bool) (kmacctmgr.cpp:184)
==7545==    by 0x64A115: KMAcctMgr::singleCheckMail(KMAccount*, bool) (kmacctmgr.cpp:118)
==7545==    by 0x64ABE1: KMAcctMgr::checkMail(bool) (kmacctmgr.cpp:286)
==7545==    by 0x733FD6: KMMainWidget::slotCheckMail() (kmmainwidget.cpp:735)
==7545==    by 0x731E3C: KMMainWidget::readConfig() (kmmainwidget.cpp:331)
==7545==    by 0x730B6E: KMMainWidget::KMMainWidget(QWidget*, char const*, KActionCollection*, KConfig*) (kmmainwidget.cpp:118)
==7545==    by 0x5686C7: KMMainWin::KMMainWin(QWidget*) (kmmainwin.cpp:19)
==7545==    by 0x6CE8A8: KMKernel::openReader() (kmkernel.cpp:198)
==7545==    by 0x6D35FF: KMKernel::action(bool, bool, QString const&, QString const&, QString const&, QString const&, QString const&, KURL const&, KURL::List const&) (kmkernel.cpp:1262)
==7545==    by 0x804C5BD: KMailApplication::newInstance() (main.cpp:159)
==7545==    by 0x678DF65: KUniqueApplication::newInstanceNoFork() (kuniqueapplication.cpp:343)
==7545==    by 0x678E5C8: KUniqueApplication::qt_invoke(int, QUObject*) (kuniqueapplication.moc:87)
==7545==    Address 0x424589B4 is 44 bytes inside a block of size 48 free'd
==7545==    at 0xA41DFD: __builtin_delete (vg_replace_malloc.c:244)
==7545==    by 0xA41E28: operator delete(void*) (vg_replace_malloc.c:253)
==7545==    by 0x170E6C4: QGuardedPtrPrivate::~QGuardedPtrPrivate() (in /usr/local/qt-copy/lib/libqt-mt.so.3.2.2)
==7545==    by 0x59057D: QGuardedPtr<KMFolder>::deref() (qguardedptr.h:125)
==7545==    by 0x590532: QGuardedPtr<KMFolder>::~QGuardedPtr() (qguardedptr.h:82)
==7545==    by 0x590479: QValueListNode<QGuardedPtr<KMFolder> >::~QValueListNode() (qvaluelist.h:280)
==7545==    by 0x5AAA7C: QValueListPrivate<QGuardedPtr<KMFolder> >::clear() (qvaluelist.h:386)
==7545==    by 0x5A90BC: QValueList<QGuardedPtr<KMFolder> >::clear() (qvaluelist.h:590)
==7545==    by 0x69B82B: KMAcctImap::slotUpdateFolderList() (kmacctimap.cpp:346)
==7545==    by 0x69BD1A: KMAcctImap::qt_invoke(int, QUObject*) (kmacctimap.moc:113)
==7545==    by 0x17441A5: QObject::activate_signal(QConnectionList*, QUObject*) (in /usr/local/qt-copy/lib/libqt-mt.so.3.2.2)
==7545==    by 0x1744063: QObject::activate_signal(int) (in /usr/local/qt-copy/lib/libqt-mt.so.3.2.2)
==7545==    by 0x6558F1: KMFolderMgr::changed() (kmfoldermgr.moc:130)
==7545==    by 0x654E20: KMFolderMgr::contentsChanged() (kmfoldermgr.cpp:347)
==7545==

Subject: Re:  another crash in kacctimap

On Thursday 27 November 2003 09:02, Stephan Kulow wrote:

> ------- Additional Comments From coolo@kde.org  2003-11-27 09:02 -------
> This is the result of the polish jury: Germany 2 points ;)

[snip backtraces]

Thanks, Coolo, this shows that while iterating over the list of folders to 
check mail in that list is modified because a folder is no longer present on 
the server. This leads to the iterator currently in use becoming invalid and 
a crash.

Folks, what's the cleanest way to fix this? Is there a safer way to work with 
QValueList iterators? maybe while(!end) and then immediate increment, before 
the list is potentially changed? Would that be safe?

Till

And it looks like #68934 is another symptom of the same problem. Different backtrace, but same issue, I would guess.

Subject: Re:  another crash in kacctimap         

> ... this shows that while iterating over the list of folders to 
> check mail in that list is modified because a folder is no longer present on 
> the server. This leads to the iterator currently in use becoming invalid and 
> a crash.

Hang on, this happends on normal basis usage of email client. No folders 
are removed/added/moved whatsoever. Guy has 3 accounts, one pop3 and 2 
imap ones. Pop3 one is pasword proteced, and if he won't unlock it on 
kmail start, but later it won't crash (well, it still does but more 
randomly). That is all information i could get from him.

--
GJ 

Subject: kdepim/kmail

CVS commit by burghard: 

This should fix 69004, 69157, 68992
Can anybody confirm this?

CCMAIL: 69004@bugs.kde.org
CCMAIL: 69157@bugs.kde.org
CCMAIL: 68992@bugs.kde.org


  M +5 -6      kmacctimap.cpp   1.146


--- kdepim/kmail/kmacctimap.cpp  #1.145:1.146
@@ -207,8 +207,5 @@ void KMAcctImap::ignoreJobsForMessage( K
     {
       if ( job->mJob )
-      {
-        job->mJob->disconnect();
         removeJob( job->mJob );
-      }
       mJobList.remove( job );
       delete job;
@@ -297,4 +294,5 @@ void KMAcctImap::processNewMail(bool int
     }
   }
+  bool gotError = false;
   // then check for new mails
   for (it = mMailCheckFolders.begin(); it != mMailCheckFolders.end(); it++)
@@ -321,11 +319,12 @@ void KMAcctImap::processNewMail(bool int
             // there was an error so cancel
             mCountRemainChecks--;
-            // just in case the folder is gone
-            slotUpdateFolderList();
-          }
+            gotError = true;
         }
       }
     }
   }
+  } // end for
+  if ( gotError )
+    slotUpdateFolderList();
 }
 

it still segfaults *g*

germany -3 points,  hehe

backtrace:
[New Thread 16384 (LWP 32609)]
0x030eebb8 in waitpid () from /lib/libpthread.so.0
#0  0x030eebb8 in waitpid () from /lib/libpthread.so.0
#1  0x06881d00 in __JCR_LIST__ () from /opt/kde/lib/libkdecore.so.4
#2  0x067bd748 in KCrash::defaultCrashHandler(int) (sig=11) at kcrash.cpp:246
#3  0x030ed7f5 in __pthread_sighandler () from /lib/libpthread.so.0
#4  <signal handler called>
#5  0x01812180 in QGList::QGList(QGList const&) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#6  0x00c6e898 in QPtrList (this=0xbffa0f90, l=@0xa6526c8) at qptrlist.h:69
#7  0x00d07ef0 in KMail::FolderJob::msgList() const (this=0xa6526a0)
    at folderjob.cpp:86
#8  0x00c54b0a in KMAcctImap::ignoreJobsForFolder(KMFolder*) (this=0x9ff9cd8,
    folder=0x9fdd198) at kmacctimap.cpp:224
#9  0x00c65a8f in KMFolderImap::close(bool) (this=0x9fdd198, aForced=true)
    at kmfolderimap.cpp:88
#10 0x00bd257b in KMFolder::remove() (this=0x9fdd198) at kmfolder.cpp:721
#11 0x00c0e5bb in KMFolderMgr::removeFolderAux(KMFolder*) (this=0x9fdf3b0,
    aFolder=0x9fdd198) at kmfoldermgr.cpp:314
#12 0x00c0e456 in KMFolderMgr::remove(KMFolder*) (this=0x9fdf3b0,
    aFolder=0x9fdd198) at kmfoldermgr.cpp:300
#13 0x00c67cae in KMFolderImap::slotListResult(QStringList, QStringList, QStringList, KMail::ImapAccountBase::jobData const&) (this=0x9fda160, mSubfolderNames=
        {<QValueList<QString>> = {sh = 0xa0e4248}, <No data fields>},
    mSubfolderPaths=
        {<QValueList<QString>> = {sh = 0xa0e4220}, <No data fields>},
    mSubfolderMimeTypes=
        {<QValueList<QString>> = {sh = 0xa240d28}, <No data fields>},
    jobData=@0xa5e3328) at kmfolderimap.cpp:553
#14 0x00c6e0f0 in KMFolderImap::qt_invoke(int, QUObject*) (this=0x9fda160,
    _id=18, _o=0xbffa12e0) at kmfolderimap.moc:272
#15 0x015701a6 in QObject::activate_signal(QConnectionList*, QUObject*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#16 0x00c4e4b5 in KMail::ImapAccountBase::receivedFolders(QStringList, QStringList, QStringList, KMail::ImapAccountBase::jobData const&) (this=0x9ff9cd8, t0=
        {<QValueList<QString>> = {sh = 0xa0e4248}, <No data fields>}, t1=
        {<QValueList<QString>> = {sh = 0xa0e4220}, <No data fields>}, t2=
        {<QValueList<QString>> = {sh = 0xa240d28}, <No data fields>},
    t3=@0xa5e3328) at imapaccountbase.moc:164
#17 0x00c4d198 in KMail::ImapAccountBase::slotListResult(KIO::Job*) (
    this=0x9ff9cd8, job=0xb399e28) at imapaccountbase.cpp:397
#18 0x00c4e6d9 in KMail::ImapAccountBase::qt_invoke(int, QUObject*) (
    this=0x9ff9cd8, _id=6, _o=0xbffa14d0) at imapaccountbase.moc:185
#19 0x00c55734 in KMAcctImap::qt_invoke(int, QUObject*) (this=0x9ff9cd8,
    _id=6, _o=0xbffa14d0) at kmacctimap.moc:120
#20 0x0157022c in QObject::activate_signal(QConnectionList*, QUObject*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#21 0x01136547 in KIO::Job::result(KIO::Job*) (this=0xb399e28, t0=0xb399e28)
    at jobclasses.moc:156
#22 0x0112470c in KIO::Job::emitResult() (this=0xb399e28) at job.cpp:206
#23 0x01125e71 in KIO::SimpleJob::slotFinished() (this=0xb399e28)
    at job.cpp:530
#24 0x0112bf63 in KIO::ListJob::slotFinished() (this=0xb399e28) at job.cpp:1815
#25 0x01139a5d in KIO::ListJob::qt_invoke(int, QUObject*) (this=0xb399e28,
    _id=16, _o=0xbffa1750) at jobclasses.moc:1370
#26 0x0157022c in QObject::activate_signal(QConnectionList*, QUObject*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#27 0x01570064 in QObject::activate_signal(int) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#28 0x0111762f in KIO::SlaveInterface::finished() (this=0xb292520)
    at slaveinterface.moc:226
#29 0x01115df8 in KIO::SlaveInterface::dispatch(int, QMemArray<char> const&) (
    this=0xb292520, _cmd=104, rawdata=@0xbffa1930) at slaveinterface.cpp:255
#30 0x01115a9a in KIO::SlaveInterface::dispatch() (this=0xb292520)
    at slaveinterface.cpp:191
#31 0x011136cb in KIO::Slave::gotInput() (this=0xb292520) at slave.cpp:294
#32 0x01114fe7 in KIO::Slave::qt_invoke(int, QUObject*) (this=0xb292520,
    _id=4, _o=0xbffa1a50) at slave.moc:113
#33 0x0157022c in QObject::activate_signal(QConnectionList*, QUObject*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#34 0x0157038d in QObject::activate_signal(int, int) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#35 0x0188e9e2 in QSocketNotifier::activated(int) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#36 0x0158c1b0 in QSocketNotifier::event(QEvent*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#37 0x01514a6f in QApplication::internalNotify(QObject*, QEvent*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#38 0x0151407b in QApplication::notify(QObject*, QEvent*) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#39 0x06732e63 in KApplication::notify(QObject*, QEvent*) (this=0xbffa2000,
    receiver=0xb3956c0, event=0xbffa1d70) at kapplication.cpp:509
#40 0x015047aa in QEventLoop::activateSocketNotifiers() ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#41 0x014c1148 in QEventLoop::processEvents(unsigned) ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#42 0x01528248 in QEventLoop::enterLoop() ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#43 0x015280f8 in QEventLoop::exec() ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#44 0x01514cc1 in QApplication::exec() ()
   from /usr/local/qt-copy/lib/libqt-mt.so.3
#45 0x0804c7d4 in main (argc=7, argv=0xbffa2174) at main.cpp:212

Subject: kdepim/kmail

CVS commit by burghard: 

A little more security, this should also fix the very latest crash of #69004 and get germany 1 point back.

CCMAIL: 69004@bugs.kde.org


  M +1 -4      kmacctimap.cpp   1.147


--- kdepim/kmail/kmacctimap.cpp  #1.146:1.147
@@ -222,11 +222,8 @@ void KMAcctImap::ignoreJobsForFolder( KM
     ImapJob *job = it.current();
     ++it;
-    if ( job->msgList().first()->parent() == folder ) 
+    if ( !job->msgList().isEmpty() && job->msgList().first()->parent() == folder ) 
     {
       if ( job->mJob )
-      {
-        job->mJob->disconnect();
         removeJob( job->mJob );
-      }
       mJobList.remove( job );
       delete job;

*** Bug 69671 has been marked as a duplicate of this bug. ***

*** Bug 69673 has been marked as a duplicate of this bug. ***

Created attachment 3617 [details]
vagrind logs

just few valgrind logs. Those might help you guys.

Do you still encounter this crash?

Subject: Re:  another crash in kacctimap

Nope, at least not in the same area :-)
i think it is safe to clos this bug

As requested