306801 - convert canvas to Thebes

Status: RESOLVED FIXED

(Core :: Graphics: Canvas2D, defect, P2)

Description

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

the <canvas> impl needs to be converted to use thebes instead of cairo directly,
so that eventually we can stop exporting cairo symbols from gkthebes.

Comment 1

[Stuart Parmenter]

woot woot. lets land this.

Comment 2

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

"this" the magic patch that's not here? :)  I threw away my first thebes conversion because it was done before the recent canvas work for 1.8/1.9; I have to redo it again.  Canvas is ok as-is right now, but it's something I want to do for 1.9.

Comment 3

[Eric Butler [:ebutler]]

Attachment: ver 0.1

Rough initial patch, getting rid of all direct cairo calls as possible, as well as cleaning up a few functions here and there. There are still some cairo calls left in due to lack of thebes support. In particular, there is no way to copy a path and restore and old path.

One neat thing this also fixes is the problem that canvas wasn't getting CMS since the calls didn't go through thebes.

Comment 4

[Eric Butler [:ebutler]]

Attachment: ver 0.2

Improved patch; now all Cairo calls are removed from Canvas.

Comment 5

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Comment on attachment 330297 [details] [diff] [review]
ver 0.2

This looks fine, just a few nits:

11:48 < vlad> ebutler: in nsCanvasPattern, get rid of the mData member
11:49 < vlad> ebutler: for enum Style, uncomment STYLE_MAX and get rid of the VC6 sucks #define -- we're not building with VC6 any more

There are two places where you check !mSurface.get() -- just check !mSurface.

In your DrawRect function, add spaces around the == used in the condition.

In the "XXX cairo bug workaround", there's a method on mThebes that does that same thing; should use it instead -- mThebes->UpdateSurfaceClip();

Looks great otherwise!

Comment 6

[Eric Butler [:ebutler]]

Attachment: ver 1.0

Changed a few things:
fixed nits.
Changed functions that take a enum Style to actually take Style instead of PRInt32.
Replaced some of the NS_ADDREFs with forgets.

Comment 7

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Comment on attachment 330872 [details] [diff] [review]
ver 1.0

Looks good, thanks!

Comment 8

[Rob Arnold [:robarnold]]

Pushed to mozilla-central:
http://hg.mozilla.org/mozilla-central/index.cgi/rev/4e627e64d650