403244 - crash [@ Tokenizer::tokenizeHeaders]

Status: VERIFIED FIXED

(MailNews Core :: Filters, defect)

Description

[Wayne Mery (:wsmwk)]

topcrash (top 10)

All OS.  talkbacks below are 2.0.0.6.  not found in topcrasher lists for .0.5, .0.4, or any 1.5.x

The majority, but not all, mention closing thunderbird

TB37884662 Exiting after recieving 55th 7MB email.
TB37837104 just turned on the computer and clicked the icon
TB37828268 just initiating Thunderbird,and it failed to load
TB37828748 Looking at my email.
TB37749364 Closing the application

TB37749364 
Tokenizer::tokenizeHeaders  [mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp, line 365]
TokenStreamListener::ProcessHeaders  [mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp, line 745]
nsMimeHtmlDisplayEmitter::WriteHTMLHeaders  [mozilla/mailnews/mime/emitters/src/nsMimeHtmlEmitter.cpp, line 281]
nsMimeHtmlDisplayEmitter::EndHeader  [mozilla/mailnews/mime/emitters/src/nsMimeHtmlEmitter.cpp, line 382]
MimeMessage_parse_line  [mozilla/mailnews/mime/src/mimemsg.cpp, line 286]
convert_and_send_buffer  [mozilla/mailnews/mime/src/mimebuf.cpp, line 185]
mime_LineBuffer  [mozilla/mailnews/mime/src/mimebuf.cpp, line 273]
MimeObject_parse_buffer  [mozilla/mailnews/mime/src/mimeobj.cpp, line 284]
nsStreamConverter::OnDataAvailable  [mozilla/mailnews/mime/src/nsStreamConverter.cpp, line 911]
nsStreamListenerTee::OnDataAvailable  [mozilla/netwerk/base/src/nsStreamListenerTee.cpp, line 97]
nsOnDataAvailableEvent0::HandleEvent  [mozilla/netwerk/base/src/nsAsyncStreamListener.cpp, line 425]
nsStreamListenerEvent0::HandlePLEvent  [mozilla/netwerk/base/src/nsAsyncStreamListener.cpp, line 122]
0x778b0c24
nsBindingManager::QueryInterface  [mozilla/content/xbl/src/nsBindingManager.cpp, line 305]
0x4e8d0388
0xffe512e8


TB37884662 
Tokenizer::tokenizeHeaders  [mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp, line 365]
TokenStreamListener::ProcessHeaders  [mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp, line 745]
nsMimeHtmlDisplayEmitter::WriteHTMLHeaders  [mozilla/mailnews/mime/emitters/src/nsMimeHtmlEmitter.cpp, line 281]
nsMimeHtmlDisplayEmitter::EndHeader  [mozilla/mailnews/mime/emitters/src/nsMimeHtmlEmitter.cpp, line 382]
MimeMessage_parse_line  [mozilla/mailnews/mime/src/mimemsg.cpp, line 286]
convert_and_send_buffer  [mozilla/mailnews/mime/src/mimebuf.cpp, line 185]
mime_LineBuffer  [mozilla/mailnews/mime/src/mimebuf.cpp, line 273]
MimeObject_parse_buffer  [mozilla/mailnews/mime/src/mimeobj.cpp, line 284]
nsStreamConverter::OnDataAvailable  [mozilla/mailnews/mime/src/nsStreamConverter.cpp, line 911]
nsMailboxProtocol::ReadMessageResponse  [mozilla/mailnews/local/src/nsMailboxProtocol.cpp, line 603]
nsMailboxProtocol::ProcessProtocolState  [mozilla/mailnews/local/src/nsMailboxProtocol.cpp, line 701]
nsMsgProtocol::OnDataAvailable  [mozilla/mailnews/base/util/nsMsgProtocol.cpp, line 350]
nsInputStreamPump::OnStateTransfer  [mozilla/netwerk/base/src/nsInputStreamPump.cpp, line 494]
nsInputStreamPump::OnInputStreamReady  [mozilla/netwerk/base/src/nsInputStreamPump.cpp, line 398]
nsOutputStreamReadyEvent::EventHandler  [mozilla/xpcom/io/nsStreamUtils.cpp, line 121]
0x778b0c24
nsBidiPresUtils::ReorderUnicodeText  [mozilla/layout/base/nsBidiPresUtils.cpp, line 1241]
0x5b5ec38b

Comment 1

[Wayne Mery (:wsmwk)]

>not found in topcrasher lists for .0.5, .0.4, or any 1.5.x

this is now found to be incorrect - appears for 1.5.0.13 and 2.x updates prior to 0.6

caused by bayesian?

Comment 2

[Magnus Melin [:mkmelin]]

Crash is here, must be mimehdrpar that is null: http://bonsai.mozilla.org/cvsblame.cgi?file=/mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp&mark=365&rev=MOZILLA_1_8_BRANCH#365

Patch coming up.

Comment 3

[Magnus Melin [:mkmelin]]

Attachment: proposed fix

Another possibility would be to check mimehdrpar later, but i don't know if it's worth it.

Comment 4

[David :Bienvenu]

Comment on attachment 295969 [details] [diff] [review]
proposed fix

thx, Magnus - can you just use NS_ENSURE_SUCCESS(rv, rv); ?

It makes sense that this could happen at shutdown.

Comment 5

[Magnus Melin [:mkmelin]]

Attachment: proposed fix

Actually, that doesn't work since it's a void method.
Looking at it some more, I think we might as well move mimehdrpar nearer to where it's used since it's used only for one case.

Comment 6

[David :Bienvenu]

Comment on attachment 296000 [details] [diff] [review]
proposed fix

sure, that's fine. There is also a form of NS_ENSURE_SUCCESS that works for void functions (iirc, NS_ENSURE_SUCCESS(rv,))

Comment 7

[Magnus Melin [:mkmelin]]

Seems so, would never have guessed NS_ENSURE_SUCCESS(rv, ) works! But it's very little used - only ~10 hits tree wide. 

Checking in mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp;
/cvsroot/mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp,v  <--  nsBayesianFilter.cpp
new revision: 1.70; previous revision: 1.69
done

->FIXED

Comment 8

[Magnus Melin [:mkmelin]]

Comment on attachment 296000 [details] [diff] [review]
proposed fix

Requesting branch approval for this crasher.

Comment 9

[Daniel Veditz [:dveditz]]

Comment on attachment 296000 [details] [diff] [review]
proposed fix

approved for 1.8.1.12, a=dveditz for release-drivers

Comment 10

[Magnus Melin [:mkmelin]]

MOZILLA_1_8_BRANCH
Checking in mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp;
/cvsroot/mozilla/mailnews/extensions/bayesian-spam-filter/src/nsBayesianFilter.cpp,v  <--  nsBayesianFilter.cpp
new revision: 1.50.4.6; previous revision: 1.50.4.5

Comment 11

[Samuel Sidler (old account; do not CC)]

Is there a good test we can run to verify this? Or should we just wait for release and confirm that the topcrash isn't as popular.

The most recent build for which there's a talkback report for is MozillaOrgThunderbird2Win322007121604 but I'm not sure we should call this verified until Thunderbird 2.0.0.12 has been released and the topcrash is not as popular.

Comment 12

[Wayne Mery (:wsmwk)]

(In reply to comment #11)
> Is there a good test we can run to verify this? Or should we just wait for
> release and confirm that the topcrash isn't as popular.

ss, it will be possible when bug Bug 499926 (need TB20022 topcrash list) gets fixed. :)

Comment 13

[Wayne Mery (:wsmwk)]

v.fixed - this is no longer in topcrash list as of TB20022