378330 - Crashes after loading page unsuccesfully

Status: RESOLVED DUPLICATE of bug 291643

(Toolkit :: Downloads API, defect)

Description

[Erik den Houter]

User-Agent:       Mozilla/5.0 (Windows; U; Win98; nl; rv:1.8.1.3) Gecko/20070309 Firefox/2.0.0.3
Build Identifier: Mozilla/5.0 (Windows; U; Win98; nl; rv:1.8.1.3) Gecko/20070309 Firefox/2.0.0.3

Was busy configuring Thunderbird for Webmail (Live.nl, unsupported domain ?).
Had had a "undefined is an unsupported site" error, and had entered
that frase in google search, and found somehow
http://www.ndfr.net/sitemap.xml

Page loads (progress bar fills slow to the end) but hourglass stays (Firefox hangs, nothing of the page is shown), and all else freezes (except mousemovement and taskbar). Program Manager says: "Firefox doesn't react". 
When i close Firefox down there after a minute, a crash occurs (Norton crashguard comes up) and Norton says "damaged beyond repair, restart the program")
Same happens if i evaluate page in DOM inspector, at same memory location.
Happened ten times, every load of that page.
! No Talkback comes up !

Norton Utilities 2000 crashguard:

Windows 98-versie 4.10.2222

FIREFOX.EXE heeft fout #c0000005 veroorzaakt in (Onbekend) op adres 017f:00d8fe2d 
Registers: 
EAX=bff719b8 CS=017f EIP=00d8fe2d EFLGS=00010256 
EBX=00000000 SS=0187 ESP=00d8fb38 EBP=00d8fb30 
ECX=00000000 DS=0187 ESI=00000003 FS=586f 
EDX=00008b6e ES=0187 EDI=0000060c GS=0000 
Bytes in CS:EIP: 
32 96 00 70 45 ab 00 00 00 00 00 78 ff d8 00 60 
Stapel-dump: 
2e00d8ff bff7186d 8b4e4727 00000000 58af8b4e 0000016f 02460000 8bac0000 3caa0002 16bf0000 27370000 0000016f 273758af 8bb2016f 16df3d76 00000004





Reproducible: Always

Steps to Reproduce:
1.load page (hangs)
2.Shut down Firefox in program manager (Ctrl+Alt+Del)
3.
Actual Results:  
Program must be shut down (used Norton) but session can go on, as you see
as i am reporting this now after ten crashes.


Have the Dutch versions.
Don't know what latest Dutch version is, so don't know if English
have later/better one.

Have Firefox/Thunderbird since 5 weeks, and about five different Talkbacks are reported, all starting with an aditional Norton Crashguard report.

Comment 1

[Steve England [:stevee]]

Dupe of bug 291643?

Comment 2

[Ria Klaassen (not reading all bugmail)]

Stack (Process Explorer) of this is:

ntoskrnl.exe!KiDispatchInterrupt+0x7f
ntoskrnl.exe!ExFreePool+0x447
win32k.sys+0x2acb
win32k.sys+0x518d
ntdll.dll!KiFastSystemCallRet
firefox.exe!Java_netscape_javascript_JSObject_initClass+0x9336b

Comment 3

[Ria Klaassen (not reading all bugmail)]

Crash.exe makes this:

Incident ID: 31430961
Stack Signature	nsLineBox::IndexOf 363f6388
Product ID	FirefoxTrunk
Build ID	2007042204
Trigger Time	2007-04-22 07:49:05.0
Platform	Win32
Operating System	Windows NT 5.1 build 2600
Module	firefox.exe + (0036b255)
URL visited	
User Comments	
Since Last Crash	99 sec
Total Uptime	183 sec
Trigger Reason	Breakpoint hit
Source File, Line No.	e:\builds\tinderbox\fx-trunk\winnt_5.2_depend\mozilla\layout\generic\nslinebox.cpp, line 274
Stack Trace 	
nsLineBox::IndexOf  [mozilla/layout/generic/nslinebox.cpp, line 274]
PresShell::FrameNeedsReflow  [mozilla/layout/base/nspresshell.cpp, line 3451]
nsCSSFrameConstructor::StyleChangeReflow  [mozilla/layout/base/nscssframeconstructor.cpp, line 9805]
nsCSSFrameConstructor::RestyleElement  [mozilla/layout/base/nscssframeconstructor.cpp, line 9978]
0x035838d0

Comment 4

[Erik den Houter]

Don't know your way of communicating here (as with Ria Klaassen reacting and so)
Also have not all (enough) knowledge of stack(dump) to assist fully in this.

It seems indeed same (sort) as bug 291643

What i today did is checking the memory use, and at loading the page, indeed the memory available is consumed at a rate of 1 MB per 8 sec, continuing for long time, havn't seen the end yet. After half an hour it starts eating virtual memory.    (P3 450 Mhz 284 MB 768MB virt)

What i today also saw:
I started to load that page in a second instance Firefox, but after some system monitoring i tried to get focus on the first instance (with this bugzilla bug-page), but couldn't get it; when pressed any of the two in taskbar, always the one loading that page was activated. 
I've also tried the old DrWatson to get a larger stack dump, but nothing.
(Also don't know if you can use that sort of action)
System monitoring gives an almost continues flow of "page errors" at around 13, while slowly the "interchangable memory" rises, and "locked memory" decreases.

What i also did:
load page with IE (takes half minute), send it with IE (trough file menu > send email > Thunderbird to my second email adress, and there opened the page (in same Thunderbird) as email.
EXACT same happens, although after few minutes a box tells me something like:
"script is busy or stuck;  end loading script ?" Clicking this box away doesn't stop the memory from drying out slowly, the proces goes on.
Closing Thunderbird down here doesn't end in crash.

Don't know if i coul'd further be of any assistance here.

Comment 5

[Ria Klaassen (not reading all bugmail)]

Hm, I see now that if I wait long enough, I see the site finally.

Comment 6

[Erik den Houter]

Oh no...
Have found that if i don't load Norton crash-guard, NO crash occurs. Nice...
Windows CAN close this busy Firefox without to many problems, but not Norton it seems. ("freezecheck" AND "close" both reported a crash of Firefox)

And indeed i found the page is there as long as you choose to wait an hour.

What now maybe is the only problem left:
Can't stop loading a huge (HUGE) page (Ha, when i save it it is < 2 MB) without stopping Firefox, while IE can load same page in 1 minute AND shows in an instant what the first screen is, and can be stopped.
Can anyone tell if this is still seen as a bug ? Or desighn floor ? Or bad website ?
Don't have the web-experience to know how to look at this.

How to look at the discussion at bug 291643 which doesn't concludes to me, but that is there since Version 1 ?