I introduced this regression ( yesterday ) while fixing a MLK. Test case: <?xml version="1.0"?> <?xml-stylesheet href="chrome://global/skin/" type="text/css"?> <!DOCTYPE window> <window style="height: 100%; width: 100%;" xmlns:html="http://www.w3.org/TR/REC-html40" xmlns="http://www.mozilla.org/keymaster/gatekeeper/there.is.only.xul" onload="MyOnLoad();"> <!-- <html:script language="JavaScript" src="resource:/res/samples/test2.js" /> --> <html:script> function MyOnLoad() { for ( var i = 0; i < 100; i++ ) dump( "Hello World " + i + "\n" ); } </html:script> <html:div> Hello World </html:div> </window> Expected: Mozilla should display an XML error. Actual: Error not getting displayed ( Mozilla might crash too! ).

Setting M16...and nominating for nsbeta2.

Updating status whiteboard. I just reviewed and approved Harish's fix. Great work, Harish. You identified this regression yourself, assigned yourself a bug for it, fixed the problem and got it reviewed, all in the space of 1 day!

Marking as crash, and recommend for PDT approval today.

Making sure that the XML errors are handled before calling DidBuildModel(). Fixed.

Not fixed 2000-07-07-10-M17 : WinNT Saved the test code to a file and loaded it. Browser doesn't hang, but rendering stops and no error message is displayed. reopening.

Putting on [nsbeta2-] radar. Not critical to beta2. Removing crash keyword.

oops, my bad, saved this as an html file! This bug is verified fixed. (sorry Harish!) 2000-07-07-10-M17 : WinNT

and verified...