Bug 137033

Summary:	excessive cpu usage while processing malformed headers
Product:	[Unmaintained] kmail	Reporter:	Dirk Mueller <mueller>
Component:	general	Assignee:	kdepim bugs <kdepim-bugs>
Status:	RESOLVED UNMAINTAINED
Severity:	normal	CC:	debian, kollix, vkrause
Priority:	NOR
Version:	unspecified
Target Milestone:	---
Platform:	Compiled Sources
OS:	Linux
Latest Commit:		Version Fixed In:
Sentry Crash Report:
Attachments:	proposed patch

Description Dirk Mueller 2006-11-08 12:17:53 UTC

Version:            (using KDE KDE 3.5.5)
Installed from:    Compiled From Sources
OS:                Linux

Sune Vuorela reports that certain messages cause KMail to hang. I'll attach a sample.

Comment 1 Dirk Mueller 2006-11-08 12:21:28 UTC

actually I cannot attach it but it is available under http://mirror.pusling.com/badmail

Comment 2 Dirk Mueller 2006-11-08 18:08:55 UTC

further comment from Sune: 

Notice line 23 which is about 4-500 k of backslashes.
'litb' on #kde-devel talked about start writing a patch for this.

I inspected Mail/inbox/cur and removed line 23 from that file - and after that 
there was no problems with the email. It looks like a bug in the 
header-parsing functions.

Comment 3 Sune Vuorela 2006-11-08 20:15:15 UTC

Created attachment 18472 [details]
proposed patch

 
Johannes Schaub has created a patch for this

Comment 4 Johannes Schaub 2006-11-11 16:46:35 UTC

yeah, i created it and it works fine and _alot_ of faster than the old code. the email shows up within < 1s

Comment 5 Dirk Mueller 2006-11-11 23:45:07 UTC

nice! using QString might be a good solution, given that QByteArray/QCString doesn't have a reserve() method for overcoming the O(n^2) behaviour. however, the local8Bit() conversion is wrong - it should be latin1.

Comment 6 Sune Vuorela 2006-11-16 01:44:52 UTC

so dirk is saying that the patch would be fine after doing a s/local8Bit/latin1/ over the patch ?

Or have I missed something?
A official approved fix would be nice.

/Sune

Comment 7 Allen Winter 2007-03-06 01:59:39 UTC

I recall discussing this a little on irc.  But I don't remember what we decided.  Dirk, do you remember?  Ingo?

I also think Volker had a libkmime unit test for this.

Please advise if I should commit Sune's patch.

Comment 8 Sune Vuorela 2007-03-06 09:50:41 UTC

we have been shipping this patch in debian for a couple of months. We have gotten one report of some drawbacks with it, but I still consider them less important than a freezing kmail.

A user writes:
<quote>
Since version 4:3.5.5.dfsg.1-2  kmail stopped treating correctly my name in 
the header. The last à in my last name is lost (strangely enough if I write 
it not in the last position of my name in the profile it works without 
problem). Prior to that version it worked without problem. I suspect the 
culprit is
 18_optimize_splitAddressInternal.diff
</quote>

Original report is http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=405867

/Sune

Comment 9 Martin Koller 2009-08-04 15:40:06 UTC

It would be great to have such a problematic mail to test with, as Qt4 now offers QByteArray::reserve().
Note: A quick test with a mail where the displayName field has about 700000 characters did not show a problem with the function shown in this patch (KPIMUtils::EmailParseResult splitAddressInternal) but with QByteArray::replace called from QUrl::setEncodedUrl  called from KHTLPart ...
So the problem I see lies within Qt - but maybe the original problem can still be shown ?

Comment 10 Andrew Crouthamel 2018-09-04 18:19:39 UTC

Hello! Sorry to be the bearer of bad news, but this version of Kmail has been unmaintained for many years so I am closing this bug. Please try using the latest version of Kmail to see if your issue persists. If it does, please submit a new bug in "kmail2". Thank you!