535687 - [OOPP] NPN_InvalidateRect not being forwarded

Status: RESOLVED FIXED

(Core Graveyard :: Plug-ins, defect)

Description

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

This is on Ubuntu 9.04/e10s-latest/DEBUG/Flash 10.0 r32 64-bit.  I can also repro in a m-c nightly from yesterday Ubuntu 8.04/latest 32-bit flash.

STR:
(0) Set dom.ipc.plugins.enabled=true
(1) Visit the URL above.
(2) The olive in the martini glass is supposed to bounce up and down, but does not.
(3) Occlude part of the frame, e.g. by Alt-Tab'ing to another window.  The frame can be seen to animate as it begins to be occluded.
(4) Make the entire frame visible again.  The olive can be seen to be in a different position than it was before occluded.

This doesn't happen for a windowed version of this applet, and doesn't happen for in-process windowless.

This is not only a regression from IPP, but also a regression for OOPP as this applet used to work just fine windowless/out-of-process.

Comment 1

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

According to https://developer.mozilla.org/en/NPP_HandleEvent, the browser should be invoking NPP_HandleEvent on the plugin in response to NPN_InvalidateRect.  I see Invalidate coming from the plugin, but when the plugin's visible, I *don't* see HandleEvent going back across.

Comment 2

[Doug Turner (:dougt)]

for the flash on maemo, we needed to explicitly send a activate event to flash:

http://mxr.mozilla.org/mozilla-central/source/layout/generic/nsObjectFrame.cpp#5603

I am not claiming that this is the same problem in the version of flash you are using, but it might be worth a try?

If this does work, be sure not to call UpdateVisibility() on maemo other than where it is now (We want the plugin to initially be not shown, and only when SetAbsoluteScreenPosition is called switch the visibility).

Comment 3

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Oops!  http://hg.mozilla.org/projects/electrolysis/rev/d5d054a09eb7 is the culprit.

It'd be nice if we had a (ref?)test for this, but I don't know how to write it.  bent seems to be afk so I'll into the fix for this.

Comment 4

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Attachment: fix InvalidateRect regression, simplify windows special-case code

Comment 5

[Karl Tomlinson (:karlt)]

Added a mochitest:
http://hg.mozilla.org/mozilla-central/rev/b0e993f4f07e
http://hg.mozilla.org/releases/mozilla-1.9.2/rev/7f85397aa446

Comment 6

[Ben Turner (not reading bugmail, use the needinfo flag!)]

Comment on attachment 418319 [details] [diff] [review]
fix InvalidateRect regression, simplify windows special-case code

Looks fine to me.

Comment 7

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Pushed http://hg.mozilla.org/mozilla-central/rev/04d6f128da25