129827 - [RR][AltSS]Regressions from 92868: Stylesheet switching broken; theme switching crashes [@ nsHTMLReflowState::ComputeMargin]; various pages crash/broken

Status: VERIFIED FIXED

(Core :: CSS Parsing and Computation, defect)

Description

[Christopher Aillon (sabbatical, not receiving bugmail)]

This is a recent regression.  Switching alternate stylesheets no longer works on
the trunk.  Using linux 2002030822, and a current CVS build.

View > Use Stylesheet > FooSheet

Actual results:

Nothing changes immediately.  Hovering over links will change the link styles to
the new sheet.

Comment 1

[Christopher Aillon (sabbatical, not receiving bugmail)]

Works in 2002030807, but not in 2002030822.

Comment 2

[Christopher Aillon (sabbatical, not receiving bugmail)]

This appears to be caused by karnaze's checkin for bug 92868.  Backing that out
locally fixes this for me.

http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=SeaMonkeyAll&branch=HEAD&branchtype=match&dir=&file=&filetype=match&who=karnaze%25netscape.com&whotype=match&sortby=Change+Size&hours=2&date=explicit&mindate=2002%2F03%2F08+07%3A00%3A00&maxdate=2002%2F03%2F08+08%3A00%3A00&cvsroot=%2Fcvsroot

Re-assigning.

Comment 3

[Hixie (not reading bugmail)]

This is a serious regression.

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

The problem is that ENSURE_TRUE(parentContext) causes ReResolveStyleContext to
fail for the root frame.  This is a huge regression and prevents me from working
on most of the bugs I was planning to look at for 1.0 (most of which deal with
ReResolveStyleContext.)

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Well, fixing what I mentioned in my previous comment fixes the most obvious of
the problems, but there are still some significant problems remaining.

Also, could you please not write lines longer than 80 characters?

Comment 6

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I'm presuming that this regression is also what's causing theme switching to crash.

Comment 7

[Hixie (not reading bugmail)]

Should we back out the patch in question until these issues are resolved?

Comment 8

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I suspect that bug 129850 and bug 129866 may be duplicates of this bug.

Comment 9

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

*** Bug 129850 has been marked as a duplicate of this bug. ***

Comment 10

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

*** Bug 129864 has been marked as a duplicate of this bug. ***

Comment 11

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

*** Bug 129866 has been marked as a duplicate of this bug. ***

Comment 12

[Boris Zbarsky [:bzbarsky]]

upgrading severity due to crashes

Comment 13

[Mike Shaver (:shaver -- probably not reading bugmail closely)]

The regressive checkin was backed out.  Resolving this as FIXED (already
reopened 92868).

Comment 14

[Boris Zbarsky [:bzbarsky]]

*** Bug 129800 has been marked as a duplicate of this bug. ***

Comment 15

[karnaze (gone)]

dbaron, it would be really helpful if you could give some examples of what you 
mean by "but there are still some significant problems remaining" in comment #5. 

Comment 16

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

When I tried switching stylesheets on http://www.people.fas.harvard.edu/~dbaron/
, the background color changed as it should, but the text color for some of the
text didn't.

Comment 17

[Jacob Kjome]

David,

Are you sure you are not just seeing bug 63863?  Back on 2002/2/15, you said 
you were working on a patch and said that you should "...probably finish it 
sometime...".  If you have the time, please do.  It is a huge problem when 
using style switching!

later,

Jake

Comment 18

[karnaze (gone)]

With the ENSURE_TRUE(parentContext) removed, switching style contexts on
http://www.people.fas.harvard.edu/~dbaron/ looked ok to me, but I'm not as 
familar with that page, so I could be missing something.

Comment 19

[Jay Patel [:jay]]

Just a note...I noticed that during the 2 days this crash appeared before the
checkin was backed out, the nsBox::GetBorder crash mentioned in bug 117088 also
showed up again.  Coincidence or related?  Either way, this crash is gone
according to the latest Talkback data.  Marking verified.  

Comment 20

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

Yeah, this probably caused the nsBox::GetBorder, FlushMiscWidgetInfo, and
nsCSSFrameConstructor::PaintBackgroundWithSC signatures in talkback as well.