Tested 3-17-99 Win32 build under Windows 95J and Winnt 4.0J. When you go to above URL, Apprunner.exe crashes. This page uses JavaScript, but it should not crash.

I don't believe this is a I18N bug. Here is the stack trace in window nsDocument::ContentAppended(nsDocument * const 0x00e46d40, nsIContent * 0x00d8389c, int 0x00000005) line 1236 nsHTMLDocument::ContentAppended(nsHTMLDocument * const 0x00e46d40, nsIContent * 0x00d8389c, int 0x00000005) line 691 HTMLContentSink::WillInterrupt(HTMLContentSink * const 0x00e4bec0) line 1502 CNavDTD::WillInterruptParse(CNavDTD * const 0x00d65f30) line 2737 + 27 bytes nsParser::ResumeParse(nsIDTD * 0x00000000) line 762 nsParser::OnDataAvailable(nsParser * const 0x00e473f4, nsIURL * 0x00e45510, nsIInputStream * 0x00e43aa0, unsigned int 0x00001558) line 964 + 17 bytes nsDocumentBindInfo::OnDataAvailable(nsDocumentBindInfo * const 0x00e45490, nsIURL * 0x00e45510, nsIInputStream * 0x00e43aa0, unsigned int 0x00001558) line 1783 + 24 bytes OnDataAvailableProxyEvent::HandleEvent(OnDataAvailableProxyEvent * const 0x00d6f120) line 632 StreamListenerProxyEvent::HandlePLEvent(PLEvent * 0x00d6f124) line 471 + 12 bytes PL_HandleEvent(PLEvent * 0x00d6f124) line 476 + 10 bytes PL_ProcessPendingEvents(PLEventQueue * 0x00b9ed70) line 437 + 9 bytes _md_EventReceiverProc(void * 0x002a168c, unsigned int 0x0000c0f9, unsigned int 0x00000000, long 0x00b9ed70) line 799 + 9 bytes It crash at the line : mWidget->Create(parent, r, nsnull, nsnull); and mWidget is 0x000000 that time. Reassign this to hyatt since his check in 1.39 show up the cvs blame as the last person touch that line

Handing off to RickG. This doesn't seem to have anything to do with XUL, so I'm not sure why it was assigned to me in the first place. I assume viewer would crash on this URL as well.

Kipp -- this does appear to be in nsDocument. Can you check it out?

1) it works in viewer without problem 2) There are zero lines of code in "nsDocument" that have the word "mWidget" in them. So your build must have been corrupted (frank). If you can make it crash again, please get a better stack trace (if possible) or tell us how to reproduce it! thanks!

I verified this in apr0508 build.