63234 - hang during HTML parsing

Status: VERIFIED FIXED

(Core :: DOM: HTML Parser, defect)

Description

[Michael Mahan]

Build: Mozilla/5.0 (X11; U; SunOS 5.7 sun4u; en-US; m18) Gecko/2000121809
Also hangs on build: 2000121509
Doesn't hang on older build: 2000121121

Steps to reproduce: (hangs every time)
1) > mozilla http://www.washingtonpost.com/wp-dyn/articles/A14278-2000Dec15.html
Result:
page is partially drawn, then mozilla completely hangs up.  Have to ^C to quit.

I ran gdb of the hung process to get a stack trace
Stack trace of hang:
#0  0xfd35e8a4 in nsScanner::ReadWhitespace ()
#1  0xfd3540a0 in CWhitespaceToken::Consume ()
#2  0xfd33b3b8 in nsHTMLTokenizer::ConsumeWhitespace ()
#3  0xfd33a74c in nsHTMLTokenizer::ConsumeToken ()
#4  0xfd3598c0 in nsParser::Tokenize ()
#5  0xfd358c58 in nsParser::ResumeParse ()
#6  0xfd35967c in nsParser::OnDataAvailable ()
#7  0xfd5cb62c in nsDocumentOpenInfo::OnDataAvailable ()
#8  0xfd829994 in nsFileChannel::OnDataAvailable ()
#9  0xfd7d47bc in nsOnDataAvailableEvent::HandleEvent ()
#10 0xfd7d3e80 in nsStreamListenerEvent::HandlePLEvent ()
#11 0xff2197e4 in PL_HandleEvent ()
#12 0xff219718 in PL_ProcessPendingEvents ()
#13 0xff21a550 in nsEventQueueImpl::ProcessPendingEvents ()
#14 0xfe02642c in nsAppShell::SetDispatchListener ()
#15 0xfe02613c in keysym2ucs ()
#16 0xfdfc5ea0 in g_io_unix_dispatch (source_data=0x1a0df8,
current_time=0xffbee370, user_data=0x1067d8) at giounix.c:135
#17 0xfdfc7ca0 in g_main_dispatch (dispatch_time=0xffbee370) at gmain.c:656
#18 0xfdfc853c in g_main_iterate (block=0, dispatch=1) at gmain.c:877
#19 0xfdfc8770 in g_main_run (loop=0x106748) at gmain.c:935
#20 0xfdecda08 in gtk_main () at gtkmain.c:476
#21 0xfe026a14 in nsAppShell::Run ()
#22 0xfe11e144 in nsAppShellService::Run ()
#23 0x16f14 in NS_CreateNativeAppSupport ()
#24 0x17634 in main ()

I cut the page down to a minimal page that fails.  It looks like the wierd
characters at the end of the page are causeing the problems.  I'll attach the
minimal test case

Comment 1

[Michael Mahan]

Attachment: HTML which hangs mozilla

Comment 2

[Boris Zbarsky [:bzbarsky]]

Looks like a likely duplicate of bug 63206

Comment 3

[harishd]

Attachment: proposed patch v1.1

Comment 4

[harishd]

aviationnow.com ( bug 63206 ) loads up without locking up with the patch..

Comment 5

[harishd]

*** Bug 63206 has been marked as a duplicate of this bug. ***

Comment 6

[hirata masakazu]

Is this a dupe of 62874?
There are a lot of pages that contain the invalid characters in WashingtonPost's
web site.
I have been seeing this in recent Mac builds.
Changing Platform and OS to ALL.

Comment 7

[vidur (gone)]

sr=vidur for the patch. A similar change needs to be made for various other
scanner methods.

Comment 8

[John Morrison]

*** Bug 62874 has been marked as a duplicate of this bug. ***

Comment 9

[harishd]

Fix is in....

Comment 10

[bsharma]

updated qa contact.

Comment 11

[bsharma]

Verified on:
build: 2000-03-07-10-Mtrunk
platform: WinNT

The sites "washingpost.com" and "aviationnow.com" loads fine.