629838 - Fire MozAfterPaint events when empty transactions occur

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

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

Like bug 625248 but for MozAfterPaint events. This should be easier, you just need to call NotifyDidPaintForSubtree in the right place.

You'll need to write tests though.

Comment 1

[Katelyn Gadd (:kael)]

I cannot come up with a reproduction case where trunk fails to fire MozAfterPaint for empty transactions, though I admit I don't understand the graphics stack very well at this point. Can you provide a repro, or at least a description of one?

Comment 2

[Katelyn Gadd (:kael)]

Attachment: Dispatch MozAfterPaint events during empty transactions after incrementing mozPaintCount

I've attached a patch that dispatches the MozAfterPaint event after incrementing mozPaintCount during an empty transaction. The patch includes a couple tests based on the ones for the mozPaintCount bug.

Note that this patch won't merge correctly unless my patch from #625248 has been applied.

Comment 3

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

Looks great!

+        frame->PresContext()->NotifyDidPaintForSubtree();

Just use 'presContext'

Comment 4

[Katelyn Gadd (:kael)]

Attachment: Dispatch MozAfterPaint events during empty transactions after incrementing mozPaintCount (2)

Patch updated.

Comment 5

[(no longer active)]

http://hg.mozilla.org/mozilla-central/rev/f4c949162937

Comment 6

[(no longer active)]

I backed this patch as part of this pushlog <http://hg.mozilla.org/mozilla-central/pushloghtml?changeset=7e12e3e16e6c> because of the oranges it caused <http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1296852850.1296854923.2345.gz&fulltext=1>.

I'm not sure which one of the bugs was at fault, so I just backed them all out.  The assignee needs to investigate and make sure that his patch has not been the culprit, and then re-add checkin-needed.

Comment 7

[(no longer active)]

(In reply to comment #6)
> I backed this patch as part of this pushlog
> <http://hg.mozilla.org/mozilla-central/pushloghtml?changeset=7e12e3e16e6c>
> because of the oranges it caused
> <http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1296852850.1296854923.2345.gz&fulltext=1>.
> 
> I'm not sure which one of the bugs was at fault, so I just backed them all out.
>  The assignee needs to investigate and make sure that his patch has not been
> the culprit, and then re-add checkin-needed.

Sorry, this belonged to bug 631329.

Comment 8

[David Baron :dbaron:]

Relanded all but one of the patches in that push.  For this bug, that's:
http://hg.mozilla.org/mozilla-central/rev/10ecb300665e

Comment 9

[David Baron :dbaron:]

The last four comments apparently all belonged on bug 631329.

Comment 10

[David Baron :dbaron:]

er, comment 6 through comment 8 did.