Here's the stack crawl: Calling chain using A6/R1 links Back chain ISA Caller 00000000 PPC 06042428 06A26AC0 PPC 0602A58C main+00080 06A26A70 PPC 06029DF8 nsNativeViewerApp::Run()+00034 06A26A30 PPC 05D888CC nsAppShell::Run()+00108 06A26950 PPC 05D89008 nsMacMessagePump::DoMessagePump()+001AC 06A268F0 PPC 05F71C10 LPeriodical::DevoteTimeToRepeaters(const EventRecord&)+00048 06A26890 PPC 085A9720 TimerPeriodical::SpendTime(const EventRecord&)+0005C 06A26830 PPC 085A915C TimerImpl::Fire()+0002C 06A267F0 PPC 08563088 timer_callback(nsITimer*, void*)+00028 06A267B0 PPC 05E6E5F8 net_process_net_timer_callback+0004C 06A26770 PPC 05E69448 NET_ProcessNet+00520 06A263D0 PPC 05E77FE8 net_ProcessFile+000F4 06A26380 PPC 05E775B8 net_read_file_chunk+00294 06A26330 PPC 05E990B8 stub_put_block(_NET_StreamClass*, const char*, long)+000D8 06A262E0 PPC 05A06D7C nsDocumentBindInfo::OnDataAvailable(nsIURL*, nsIInputStream*, unsigned int)+000BC 06A26280 PPC 05D472D4 nsParser::OnDataAvailable(nsIURL*, nsIInputStream*, unsigned int)+00214 06A24210 PPC 05D46994 nsParser::ResumeParse()+0005C 06A241C0 PPC 05D46B4C nsParser::BuildModel()+0011C 06A24130 PPC 05D32C28 CNavDTD::HandleToken(CToken*, nsIParser*)+00094 06A240E0 PPC 05D45080 CTokenHandler::operator()(CToken*, nsIDTD*)+00048 06A24090 PPC 05D31E04 NavDispatchTokenHandler(CToken*, nsIDTD*)+00080 06A24040 PPC 05D3367C CNavDTD::HandleEndToken(CToken*)+0017C 06A23F20 PPC 05D381D4 CNavDTD::CloseContainersTo(nsHTMLTag, int)+00080 06A23EC0 PPC 05D38034 CNavDTD::CloseContainersTo(int, nsHTMLTag, int)+ 000F8 06A23E00 PPC 05D37E58 CNavDTD::CloseContainer(const nsIParserNode&, nsHTMLTag, int)+00148 06A23DB0 PPC 05D37A44 CNavDTD::CloseFrameset(const nsIParserNode&)+0005C 06A23D70 PPC 05ACAC80 HTMLContentSink::CloseFrameset(const nsIParserNode& )+0007C 06A23D20 PPC 05ACB480 HTMLContentSink::StartLayout()+00120 06A23C90 PPC 05AB0730 PresShell::InitialReflow(int, int)+002E8 06A23B70 PPC 05B44A4C RootFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+003DC 06A23A10 PPC 05A9F588 nsContainerFrame::ReflowChild(nsIFrame*, nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+ 000B4 06A239B0 PPC 05CC4858 nsAreaFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+003DC 06A23880 PPC 05C651BC nsBlockFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+00040 06A23830 PPC 05C5CE08 nsBaseIBFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+002D0 06A22E60 PPC 05C66358 nsBlockFrame::ReflowDirtyLines(nsBlockReflowState&)+ 00038 06A22E10 PPC 05C5DF10 nsBaseIBFrame::ReflowDirtyLines(nsBlockReflowState& )+00068 06A22D80 PPC 05C5E890 nsBaseIBFrame::ReflowLine(nsBlockReflowState&, nsLineBox*, int&)+00284 06A22D00 PPC 05C5FEF8 nsBaseIBFrame::ReflowInlineFrame(nsBlockReflowState& , nsLineBox*, nsIFrame*, int&, int&)+000A8 06A22C80 PPC 05C6F958 nsInlineReflow::ReflowFrame(nsIFrame*, int, unsigned int&)+00120 06A22BF0 PPC 05C6FF08 nsInlineReflow::ReflowFrame(int, nsHTMLReflowMetrics&, unsigned int&)+00200 06A22B10 PPC 05C5778C nsHTMLFramesetFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+00040 06A22AD0 PPC 05C58618 nsHTMLFramesetFrame::Reflow(nsIPresContext&, nsFramesetDrag*, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+ 00E04 06A22930 PPC 05C56DD8 nsHTMLFramesetFrame::ReflowPlaceChild(nsIFrame*, nsIPresContext&, const nsHTMLReflowState&, nsPoint&, nsSize&, nsFramesetDrag*, nsPoint*)+004D0 06A22800 PPC 05A9F588 nsContainerFrame::ReflowChild(nsIFrame*, nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+ 000B4 06A227A0 PPC 05C5268C nsHTMLFrameOuterFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+00268 06A22680 PPC 05A9F588 nsContainerFrame::ReflowChild(nsIFrame*, nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+ 000B4 06A22620 PPC 05C54400 nsHTMLFrameInnerFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+0015C 06A224E0 PPC 05A0F11C nsWebShell::LoadURL(const unsigned short*, nsIPostData*, int, nsURLReloadType, unsigned int)+0004C Return addresses on the stack Stack Addr Frame Addr ISA Caller 06A227A8 PPC 05A9F588 nsContainerFrame::ReflowChild(nsIFrame* , nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int& )+000B4 06A226C8 68K 061C5E3E t0f+FE9DE 06A226B8 06A226B0 PPC 086603D8 nsSmallHeapChunk::GetSpaceForBlock(unsigned long)+00158 06A22688 06A22680 PPC 05C5268C nsHTMLFrameOuterFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+00268 06A22674 68K 06A2285A 06A22660 68K 061C5E3E t0f+FE9DE 06A22638 06A22630 PPC 05C6D7C8 nsHTMLReflowState::DetermineFrameType(nsIPresContext&)+00204 06A22628 06A22620 PPC 05A9F588 nsContainerFrame::ReflowChild(nsIFrame* , nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int& )+000B4 06A22624 06A22620 68K 06A2285A 06A225C4 68K 06A2285A 06A22556 PPC 001E0070 GetIndSymbol+00AE4 06A22514 06A22510 68K 06A2285A 06A224E8 06A224E0 PPC 05C54400 nsHTMLFrameInnerFrame::Reflow(nsIPresContext&, nsHTMLReflowMetrics&, const nsHTMLReflowState&, unsigned int&)+0015C 06A224A8 06A224A0 PPC 05A0F11C nsWebShell::LoadURL(const unsigned short*, nsIPostData*, int, nsURLReloadType, unsigned int)+0004C 06A22448 06A22440 PPC 05C53BBC TempMakeAbsURL(nsIContent*, nsString&, nsString&)+00110 06A22408 06A22400 PPC 085A4468 nsAutoString::~nsAutoString()+00048 06A223D8 68K 05E9523A NS_MakeAbsoluteURL(nsIURL*, const nsString&, const nsString&, nsString&)+00182 06A223B8 06A223B0 PPC 05A0F430 nsWebShell::LoadURL(const unsigned short*, const char*, nsIPostData*, int, nsURLReloadType, unsigned int)+002AC

I can't reproduce it. I tried with Monday's debug build and today's optimized build. Reassigned to <beard> for more details.

CC'd myself

It crashes about 50% of the time for me in timer_callback

Setting all current Open Critical and Major to M3

This one should be fixed by now: the timers have been rewritten by jfrancis/ sfraser.

It no longer crashes, but it does still assert like crazy. It also draws very strange when it comes up, and scrolling doesn't really work with all of those frames. I also managed to crash it while dragging the left bottom frame's scroll bar around.

Renamed "[PP] Various problems with frames" and reassigned to myself.

*** Bug 3098 has been marked as a duplicate of this bug. ***

per leger, assigning QA contacts to all open bugs without QA contacts according to list at http://bugzilla.mozilla.org/describecomponents.cgi?product=Browser

glynn - your the Frames QA Contact. How's this look on Apprunner with latest build?

Changed component to HTMLFrames, will check when I get in.

Still broken in March 9 Mac build using apprunner in ways mentioned in comments; crash is gone but scrolling the frames, especially the top right frame, results in redraw chaos.

It's next on my list of priorities. I'll work on it next week.

pierre status on this one?

I keep defering it from one week to the next. It will be Real Soon Now.

Eli, I'm re-assigning you these Mac bugs to take ownership. These are Greg's old bugs...

*** Bug 4558 has been marked as a duplicate of this bug. ***

Too bad, I can't fix it for M4: pushing to M5. A simpler case of problem with frames is described under #4570 "XML demo doesn't scroll correctly".

*** Bug 5450 has been marked as a duplicate of this bug. ***

M6. Sigh...

*** Bug 5694 has been marked as a duplicate of this bug. ***

This has got to be fixed in M5 as far as I'm concerned - it really bungs up the Mac user experience.

I can only agree. Maybe we can ask to postpone M5 on the Mac by a couple of weeks.

Surprise! I have a fix, kind of brute-force but it works... Back to M5.

Re-opening. Sorry. Unfortunately, this bug report doesn't appear to provide any indication of what the actual bug is, so I can't actually discern whether it's fixed or not. (A specific statement of what the bug is would be most appreciated; thanks.) I *can* tell that using the 19999043017 build, scrolling the top half of test #9 (the URL cited in the bug report) results in every frame in the top of the page becoming filled dozens of with horizontal lines and instances of the "frame3" text, and that this problem doesn't occur on 4.29.99 Linux or Win32. Since that's problem on the frame demo page, it falls under the bug's current description of "various problems with frames", and thus, I'm re-opening it.

Renamed "[PP] Frame scrolling is broken" from "[PP] [BLOCK] Various problems with frames". To see what has been fixed already, you can compare the rendering of test9 before and after my checkin: the result is fairly clear. Moving to M6.

This time I fixed the scrolling too. Maybe the fix will be moved to the M5 branch.

Using the 1999050422 M5 candidate build for Mac OS, this bug appears to be most definitely fixed; can't reproduce after various forms of scrolling, including with the sidebar active or docked, as well as with different frame heights and widths. Thanks!