width=0 on a table tag is handled as a special case in layout code in several places. This gives the right result, but ultimately it is the wrong approach. from Steve's original note: table attribute width=0 is treated as width=auto, not width=min Peter: do we want to differentiate between HTML width=0 and CSS width=0? Peter's response: Yeah, I'd do the fixup in the MapAttributesInto code, converting 0 to auto there. Then let the layout code respect 0 as 0, then CSS won't be affected.

I know I won't have time to get to this.

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

Moving to M7.

Moving to M9.

*** Bug 5445 has been marked as a duplicate of this bug. ***

Per bug 5445, width=0% also has the meaning of width=auto under 4.x. Currently, we are only mapping width=0 as auto. When changing the way that width=0 is handled, don't forget to make sure that 0% also triggers the 'auto' code. See also the test case for bug 5445: http://bugzilla.mozilla.org/showattachment.cgi?attach_id=464

[TESTCASE] View the attachment (#769) to compare four tables (width=0, width=0%, CSS/width=0, CSS/width=0%). Currently, width=0 and CSS/width=0 map to "auto" and width=0% and CSS/width=0% map to "min". At least, that's how it appears. Both CSS widths should map to "min" and both HTML widths should map to "auto".

Moving to M13.

mass move to m14.

Bulk moving [testcase] code to new testcase keyword. Sorry for the spam!

Fixed with latest checkin.

Testcase looks fine. Based on Karnaze's comments, verifying bug fixed.