1340717 - stylo: figure out whether we can optimize the change hint for table frames better

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect, P4)

Description

[Boris Zbarsky [:bzbarsky]]

Right now we pass the same change hint for both the table wrapper and the table.  It's not clear whether we can, or should, do better for the latter by excluding things the parent will handle.  Just passing NS_HintsNotHandledForDescendantsIn(changeHint) is NOT right, because that can produce invalid hints that fail this assert in nsStyleChangeList::AppendChange:

  MOZ_ASSERT(!(aHint & nsChangeHint_AllReflowHints) ||
             (aHint & nsChangeHint_NeedReflow),
             "Reflow hint bits set without actually asking for a reflow");

(because it will copy some reflow bits, but doesn't seem to copy NeedReflow?).

Cameron, do you have time to look at this? I can do it, but will be a bit slow; we've grown a lot of changehint complexity since I last looked at it.

Comment 1

[Cameron McCormack (:heycam)]

I wonder if NS_RemoveSubsumedHints in the bug 1301258 patches does the right thing?

Comment 2

[Boris Zbarsky [:bzbarsky]]

That seems pretty likely, yes.

Comment 3

[Boris Zbarsky [:bzbarsky]]

Note that in bug 1340723 I'm going to move this stuff to a somewhat different setup which might effectively make this bug moot.

Comment 4

[Bobby Holley (:bholley)]

Boris, is this moot now per comment 3?

Comment 5

[Boris Zbarsky [:bzbarsky]]

I think so, yes.