159524 - On certain web-sites, Mozilla quits if one tries to save the page.

Status: RESOLVED WONTFIX

(Core :: Layout: Tables, defect, P1)

Description

[Michael Graubart]

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (Macintosh; U; PPC; en-US; rv:1.1b) Gecko/20020723
BuildID:    2002072308 (trunk)

NB: This was also submitted via the Quality Feedback Agent, but as that seems to
refer to Netscape as such, I felt it worth submitting through Bugzilla also.

As soon as I tried to save the whole page to disk (xxx.htm), Mozilla quitted,
and an empty folder was saved on my desktop.  This did not happen when I used
Netscape 4.79.  Other pages allow themselves to be saved in Mozilla also.

Reproducible: Always
Steps to Reproduce:
1. Open the above page.
2. Type 'Command-S' or open File menu and choose 'Save as…'.
3. Ensure that it is the whole page that is to be saved.
4. Save.

Actual Results:  Mozilla quits.  An empty folder appears on the desktop.

Expected Results:  A file should be saved on the desktop that, when opened,
displays the page that was originally displayed at the given URL.

Mac G3 - OS 9.2.2 (Intern. Eng.).

Classic theme.

Comment 1

[Boris Zbarsky [:bzbarsky]]

What was the quality feedback agent incident id for this incident?

Comment 2

[wighta]

Mozilla/5.0 (Macintosh; U; PPC; en-US; rv:1.1b) Gecko/
Build ID 2002072508. Mac OS 9.2.2 US English. Little Mozilla 1.1 theme.

I confirm this crash after Save As 'Web Page, complete'. The crash did not occur
immediately but after I moved back to this bug report within Bugzilla & then
switched back to the site whereon I crashed into MacsBug with a Type 2 Error.
See attached stdlog file. The talkback # is TB8709319G.

If the Save As is 'Web Page, HTML only' then the save occurs normally.

Comment 3

[wighta]

Attachment: stdlog file for additional comment # 2

Comment 4

[greer]

Talkback incident from comment #2
----------------------------
Product ID MozillaTrunk
Build ID 2002072508
Operating System MacOS version 9.2.2
User Comments: had been reading a bug report on Bugzilla, went to the URL that
crashed upon save, did some saves but Moz did not crash, went back to bug report
then clicked upon forward arrow in Moz anf then fell into MacsBug. Have a stdlog
file available. The crash was a Type 2 error acording to the Finder

Stack Trace
free() [nsAllocatorManager.cp]
.__dl__FPv [CPlusPlusLevel.cp]
nsCellMap::~nsCellMap() [nsCellMap.cpp, line 983]
nsTableCellMap::~nsTableCellMap() [nsCellMap.cpp, line 113]
nsTableFrame::~nsTableFrame() [nsTableFrame.cpp, line 297]
nsFrame::Destroy() [nsFrame.cpp, line 484]
nsContainerFrame::Destroy() [nsContainerFrame.cpp, line 143]
nsTableFrame::Destroy() [nsTableFrame.cpp, line 314]
nsFrameList::DestroyFrames() [nsFrameList.cpp, line 130]
nsContainerFrame::Destroy() [nsContainerFrame.cpp, line 140]
nsTableOuterFrame::Destroy() [nsTableOuterFrame.cpp, line 82]
nsLineBox::DeleteLineList() [nsLineBox.cpp, line 311]
nsBlockFrame::Destroy() [nsBlockFrame.cpp, line 421]
nsLineBox::DeleteLineList() [nsLineBox.cpp, line 311]
nsBlockFrame::Destroy() [nsBlockFrame.cpp, line 421]
nsAreaFrame::Destroy() [nsAreaFrame.cpp, line 167]
nsFrameList::DestroyFrames() [nsFrameList.cpp, line 130]
nsContainerFrame::Destroy() [nsContainerFrame.cpp, line 140]
CanvasFrame::Destroy() [nsHTMLFrame.cpp, line 253]
nsFrameList::DestroyFrames() [nsFrameList.cpp, line 130]
nsContainerFrame::Destroy() [nsContainerFrame.cpp, line 140]
nsBoxFrame::Destroy() [nsBoxFrame.cpp, line 1244]
nsFrameList::DestroyFrames() [nsFrameList.cpp, line 130]
nsContainerFrame::Destroy() [nsContainerFrame.cpp, line 140]
nsBoxFrame::Destroy() [nsBoxFrame.cpp, line 1244]
nsGfxScrollFrame::Destroy() [nsGfxScrollFrame.cpp, line 467]
nsFrameList::DestroyFrames() [nsFrameList.cpp, line 130]
nsContainerFrame::Destroy() [nsContainerFrame.cpp, line 140]

Comment 5

[Matthias Versen [:Matti]]

-> Tables

Comment 6

[Boris Zbarsky [:bzbarsky]]

From Michael's talkbacks:

realloc() [nsAllocatorManager.cp]
JS_realloc() [jsapi.c, line 1430]
js_AllocSlot() [jsobj.c, line 1852]
js_AddScopeProperty() [jsscope.c, line 1098]
js_DefineNativeProperty() [jsobj.c, line 2082]
js_DefineProperty() [jsobj.c, line 1998]
DefineProperty() [jsapi.c, line 2060]
JS_DefineProperty() [jsapi.c, line 2149]
DefineInterfaceConstants() [nsDOMClassInfo.cpp, line 3293]
nsWindowSH::GlobalResolve() [nsDOMClassInfo.cpp, line 3424]
nsWindowSH::NewResolve() [nsDOMClassInfo.cpp, line 3699]
XPC_WN_Helper_NewResolve() [xpcwrappednativejsops.cpp, line 903]
js_LookupProperty() [jsobj.c, line 2251] 

and 

0x65707468
js_DefineProperty() [jsobj.c, line 1998]
DefineProperty() [jsapi.c, line 2060]
JS_DefineProperty() [jsapi.c, line 2149]
DefineInterfaceConstants() [nsDOMClassInfo.cpp, line 3293]
nsWindowSH::GlobalResolve() [nsDOMClassInfo.cpp, line 3424]
nsWindowSH::NewResolve() [nsDOMClassInfo.cpp, line 3699]
XPC_WN_Helper_NewResolve() [xpcwrappednativejsops.cpp, line 903]
js_LookupProperty() [jsobj.c, line 2251]
js_GetProperty() [jsobj.c, line 2455]
JS_GetProperty() [jsapi.c, line 2343]
nsDOMClassInfo::PostCreate() [nsDOMClassInfo.cpp, line 2407]
XPCWrappedNative::GetNewOrUsed() [xpcwrappednative.cpp, line 401]

These are not tables stacks at all... so something else is up here.

Comment 7

[Bartosz Wucke]

Seems Mac-characteristic. Worksforme on Linux.

Comment 8

[karnaze (gone)]

wfm on win2k. Over to dcone for mac.

Comment 9

[Carine Held]

Reporter, the URL seems to have changed and is inaccessible now. Get a "The
requested URL /~solo/cageopus.htm was not found on this server." message. Can
you update the URL? If not, I suggest we close this bug.

Comment 10

[Michael Graubart]

Yes, the web-page seems to have disappeared, and I have not had any similar
trouble with other web-sites, so the bug should be closed.

Comment 11

[Boris Zbarsky [:bzbarsky]]

um... how about we have someone who may have some clue look at this _crash_ bug
first, huh?

dbradley, sorry to dump this on your but this looks like your neck of the woods...

Comment 12

[David Bradley]

Looks like we're probably dealing with a corrupted heap. This appears to be OS 9
only, no incidents showing up on OSX. Anyone ported Valgrind to Mach? Or is
there a comparable tool? That's about the only way we're going to be able to
track this down.

Comment 13

[Simon Paquet [:sipaq]]

This bug is targeted at a Mac classic platform/OS, which is no longer supported
by mozilla.org. Please re-target it to another platform/OS if this bug applies
there as well or resolve this bug.

I will resolve this bug as WONTFIX in four weeks if no action has been taken.
To filter this and similar messages out, please filter for "mac_cla_reorg".

Comment 14

[Simon Paquet [:sipaq]]

mac classic is no longer supported by Mozilla. Marking WONTFIX.