QA/Topcrashes

In progress. I have a lot more work to do on this page.

This page is a tracking document for handling and following topcrashes on the trunk.

Current topcrashes being tracked for beta 4

data for beta4 crashes in the last week

#	bug	stack	where	status
1	bug 422018	googletoolbar.dll@0x4b2f libgoogletoolbar.dylib@0x4fbd	google toolbar? core?	trying to track down someone at google.
2, 3, 14, 17	bug 416521 (70%), bug 422024 (20%)	RtlEnterCriticalSection, free, arena_dalloc_small, ntdll.dll@0x42e7b	this is what shows up when we crash in free()	bug 422024 is related to scriptable plugin API. FreeDownLoadManager (one of the problematic extensions) is now blocked
4	tried to block IDM extension in bug 382356 but it's still showing up.	MultiByteToWideChar	internet download manager	always has to do with idmmzcc.dll Internet Download Manager - which now has been blocked
5, 15	bug 421303	jsds_ScriptHookProc		Fixed for 2008-03-11 builds
6	bug 380015	nsFrame::BoxReflow	due to Firefox 3 installs that mix in components from Firefox 2
7	bug 421217	JS_SetPrivate	silverlight plugin or scriptable plugin API	need to contact Silverlight developers
8, 13	bug 419695	js_GetGCThingTraceKind, JS_CallTracer, JS_TraceChildren	crashing on yahoo mail (maybe not all of the JS_CallTracer crashes, though?)	Fixed for 2008-03-12 builds
9	bug 391311	nsChromeRegistry::CheckForNewChrome
10	bug 418384	ClientData::GetOtlTable		has testcase (requires particular font)
11	bug 418381	HashString		looks like a start-up crash. not much to go on for further investigation. running the mini-dump might help.
12	bug 422467	piclens.dll	the piclens extension	the extension doesn't appear to be compatible, or marked as such (although it can't be installed due to not providing secure updates), and the crash is what you'd expect from using a binary extension with incompatible interface changes. Is this entirely from people overriding extension compat? -- piclens develpers contacted 3/12
16		0x0	didn't find any common patterns in sample of ten; this is mixed crashes (although about half the stacks are useless) -David Baron
18	bug 418382	UniscribeItem::SaveGlyphs(gfxTextRun*)	core
1 on mac		nsObjCExceptionLogAbort
19 (2 on mac)	bug 421294	DocumentViewerImpl::GetCopyable		fixed for 2008-03-07 builds
6 on mac		libobjc.A.dylib@0x146e8 via nsMenuX::RemoveAll
11 on mac	bug 422546	nsEditor::nsEditor		fix in hand

Current topcrashes being tracked for beta 3

Topcrashers for beta3 are currently considered "old". We need to re-track for beta4. I'm using this "old" data as sample data to see if what format works.

#	bug	stack	where	status
4	405357	jpinscp.dll@0xcf45	java plugin
6	418378	nsGlobalWindow::SaveWindowState(nsISupports**)	core
7	391311	nsChromeRegistry::CheckForNewChrome	core
9	418382	UniscribeItem::SaveGlyphs(gfxTextRun*)	core
11	418381	HashString(nsAString_internal const&)	core
14	418379	nsNavHistoryFolderResultNode::FindChildById(__int64, unsigned int*)	core
26	418384	ClientData::GetOtlTable(long, unsigned char const**, unsigned long*)	core
37	380015	nsFrame::BoxReflow(nsBoxLayoutState&, nsPresContext*, nsHTMLReflowMetrics&, nsIRenderingContext*, int, int, int, int, int)	core

What do I do?

So, someone CCed you to a topcrash bug and asked for your help. What should you do?

There are several steps that developers and QA need to go through to help fix a crash. As a developer, we're specifically asking that you comment in the bug when you've completed all the steps and specifically state which steps you completed.