276154 - position: fixed; breaks on layout re-calculations

Status: RESOLVED DUPLICATE of bug 268111

(Core :: Layout, defect)

Description

[Jonathan Stanley]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8a6)
Gecko/20041227
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.2; en-US; rv:1.8a6)
Gecko/20041227

Fixed positioned elements seem to change their state to absolutely positioned
when  elements on a page change their appearance, in this case initiated by
a:hover::after inserting some generated content, a "*".

Should the page be scrolled so the fixed element is outside the viewport, this
change of state completely removes the element from showing on the page at all.

The effected element can be brought back to it's correct fixed position state by
triggering a layout state change, eg: hover over an anchor which creates
generated content.

With some non-exhaustive testing, it's not just generated content which can
cause such issues, but also changes to outline (well, -moz-outline), border,
margin, padding, font-* which are dynamically changed via :hover, :active, :focus.

This bug doesn't manifest itself in Mozilla 1.7.5 (Mozilla Suite or Firefox) nor
Mozilla 1.8 up till about Alpha4.

Comment 1

[Bernd]

please file layout bugs in the core component and not inside the application
components

Comment 2

[Martijn Wargers (dead)]

Attachment: Testcase

I've seen this bug at http://www.cleansoftware.org/
This is a testcase based from that site. This bug happens with position:fixed
elements with auto-offsets.

I don't see the bug, using:
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a5) Gecko/2004110305
I can see the bug, using:
Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8a5) Gecko/20041104
Firefox/0.9.1+ 11:53am

I think this could be a regression from fixing bug 266968.

Comment 3

[Martijn Wargers (dead)]

By the way, this is a bonsai link of that regression range:
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=all&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=2004-11-03+05%3A00%3A00&maxdate=2004-11-04+12%3A00%3A00&cvsroot=%2Fcvsroot

Boris, do you agree this is a regression from bug 266968? (if not, my apologies)

Comment 4

[Boris Zbarsky [:bzbarsky]]

Isn't this a duplicate of bug 268111?

Comment 5

[Jonathan Stanley]

I'd say it's a duplicate though steps needed to reproduce the effect is
different. It (fixed positioning) appears to have broke during the same set of
builds, between Gecko/20041103 and Gecko/20041104.

Comment 6

[Martijn Wargers (dead)]

Yes, a duplicate. Sorry about that, didn't find that previous bug.

Comment 7

[Boris Zbarsky [:bzbarsky]]

*** This bug has been marked as a duplicate of 268111 ***