991640 - Get rid of nsIWidget::GetThebesSurface and a bunch of Thebes backed gfxContexts

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Jonathan Watt [:jwatt]]

We should Moz2Dify nsIWidget::GetThebesSurface to get rid of a bunch of Thebes backed gfxContexts.

Comment 1

[Jonathan Watt [:jwatt]]

Attachment: WIP

Comment 2

[Jonathan Watt [:jwatt]]

It's only widget/gtk/nsWindow.cpp, widget/windows/nsWindow.cpp and widget/xpwidgets/PuppetWidget.h that need finished off here.

Comment 3

[Matt Woodrow (:mattwoodrow)]

It appears that the *only* caller of this function is inFlasher.cpp, and that appears to be dead code.

It can't even work for a lot of platforms where we have hardware acceleration anyway.

I think if we kill that then everything can go, except for a private function for gtk/nsWindow.

Comment 4

[Jonathan Watt [:jwatt]]

Attachment: patch

I'll do widget/gtk/nsWindow.cpp separately. I'd like to get this landed so I don't have to unbitrot it again.

Comment 5

[Jonathan Watt [:jwatt]]

Actually the DT must be cached and reused in PuppetWidget or else we get various timeouts in tests for some reason:

https://tbpl.mozilla.org/?tree=Try&rev=6beaf6199419

Comment 6

[Jonathan Watt [:jwatt]]

Attachment: patch

Comment 7

[Matt Woodrow (:mattwoodrow)]

(In reply to Jonathan Watt [:jwatt] from comment #5)
> Actually the DT must be cached and reused in PuppetWidget or else we get
> various timeouts in tests for some reason:
> 
> https://tbpl.mozilla.org/?tree=Try&rev=6beaf6199419

That's very weird. I'd have thought that path was dead code.

Comment 8

[Mike Hommey [:glandium] (high latency)]

(In reply to Matt Woodrow (:mattwoodrow) from comment #3)
> It appears that the *only* caller of this function is inFlasher.cpp, and
> that appears to be dead code.

You can exercise that code with the DOM inspector addon: https://addons.mozilla.org/en-US/firefox/addon/dom-inspector-6622/

Comment 9

[Jonathan Watt [:jwatt]]

Yes, see bug 1018324, which blocks this bug.

Comment 10

[Jonathan Watt [:jwatt]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/2e1413c075c0

Comment 11

[Jonathan Watt [:jwatt]]

Plus follow-up for Linux warning-is-error:

https://hg.mozilla.org/integration/mozilla-inbound/rev/a4859b482f44

Comment 12

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/2e1413c075c0
https://hg.mozilla.org/mozilla-central/rev/a4859b482f44