401176 - can end up with double frames in fancy nsCSSFrameConstructor::ContentRemoved codepaths

Status: RESOLVED FIXED

(Core :: Layout, defect, P2)

Description

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

Landing bug 321402 caused the reftest http://mxr.mozilla.org/seamonkey/source/layout/reftests/bugs/374038-2.xul to fail.  (It was the last-minute change (after I'd done the full reftest run) to fix the ContentRemoved codepath that broke it, although it would also have been slightly broken without that change -- although not to cause that particular test to fail.)

The problem is that we do the following:
  - call RecreateFramesForContent to reframe the image (processing the style change that resulted from ContentStatesChanged)
    - this calls ContentRemoved
      - this detects the need to destroy the wrapper and reframes its grandparent, which creates a frame for the image
    - this then calls ContentInserted, creating a *second* frame for the image

(I'm also a bit puzzled by the need for both ReinsertContent and RecreateFramesForContent, with their different brokenness, and why we set PR_TRUE for aInReinsertContent to ContentRemoved only in the former.)

Comment 1

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

(Note that this bug was present in a number of existing codepaths before bug 321402 landed.)

Comment 2

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

And when this is fixed we need to remove the random marking in layout/bugs/reftest.list:
random == 374038-2.xul 374038-2-ref.xul # bug 401176

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I will fix this when merging the patch in bug 399940

Comment 4

[Jesse Ruderman]

roc, reed ended up landing the fix for bug 399940 for you.  Did this get fixed with it anyway?

Comment 5

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Yeah, I added the fix needed here into the patch there. I'll verify as soon as I get a trunk build with that patch in it...

Comment 6

[Jesse Ruderman]

I turned the reftest back on.