122810 - view-source hangs the browser

Status: VERIFIED DUPLICATE of bug 98118

(Core :: DOM: HTML Parser, defect)

Description

[Yrjö Kari-Koskinen]

When I try to view the sources of this site, Mozilla hangs and uses CPU 100%
until it's killed. 
URL: http://www.hut.fi/u/hsivonen/HTMLSyntaxChecker-source.html

The bug was found and is reproducible with Mozilla 0.9.7 Gecko/20020121 and
Galeon 1.0.2

Comment 1

[Pete Boyd]

I see this too. Mozilla trunk 2002013003; Windows 2000

Comment 2

[Raj Bhaskar]

I see this with Mozilla 0.9.7 (build 2001-12-21-06, milestone) on Windows 95.

Comment 3

[joakim_46]

Confirming on build 2001122908 under Windows Me. Mozilla becomes unresponsive
and I have to kill it. Mark as NEW.

Comment 4

[alge]

Change severity to critical, os to all.

I got this talkback: TB 2359901H not entirely sure it's a direct consequence of
the hang but.. hey.. it crashes (:
ccing stephend for talkback.

Comment 5

[[e-mail] (gone - use [e-mail] instead)]

JS_GetClass()
XPCWrappedNative::GetWrappedNativeOfJSObject()
XPC_WN_Helper_SetProperty()
nsWebShellWindow::StoreBoundsToXUL()
nsWebShellWindow::FirePersistenceTimer()
nsTimerImpl::Process()
handleMyEvent()
PL_HandleEvent()
PL_ProcessEventsBeforeID()
processQueue()
nsVoidArray::EnumerateForwards()
nsAppShell::ProcessBeforeID()
handle_gdk_event()
libgdk-1.2.so.0 + 0x194f5 (0x403824f5)
libglib-1.2.so.0 + 0x12b90 (0x403b8b90)
libglib-1.2.so.0 + 0x130f2 (0x403b90f2)
libglib-1.2.so.0 + 0x13304 (0x403b9304)
libgtk-1.2.so.0 + 0xa8ede (0x402b7ede)
nsAppShell::Run()
nsAppShellService::Run()
main1()
main()
libc.so.6 + 0x18108 (0x404e4108) 

Comment 6

[Henri Sivonen (:hsivonen)]

Happens on OS X, too. Confirming and setting to critical, All/All. Moving to Parser.

Comment 7

[Boris Zbarsky [:bzbarsky]]

Just opened the view source to profile, killed it after a bit.  95% of the time
was spent in the nsJISx4501LineBreaker::Next function.

The entire file is a single 200-kilobyte line that we have to linebreak.  So we
spend a while doing that.  I'd wait for it to be done, but that might take a
while...

Duplicate of "nsJISx4501LineBreaker::Next is worse than O(N^2) in number of
elements in line" (and with a 200-kilobyte line that makes it _slow_).

*** This bug has been marked as a duplicate of 98118 ***

Comment 8

[Moied]

verified dup of 98118