Canvas-into-canvas drawing should respect backing store scale ratio

• Component: Canvas
• Status: RESOLVED FIXED

[Tim Horton]

When drawing between two canvases where the backing store scale ratio is not 1, using the ctx.drawImage(otherCanvas, 0, 0) form of drawImage() doesn't respect the scale ratio.

<rdar://problem/10549729>

I have a patch.

[Tim Horton]

Created attachment 125419 [details]
patch

Comment on attachment 125419 [details]
patch

Makes sense

[Tim Horton]

(In reply to comment #2)
> (From update of attachment 125419 [details])
> Makes sense

Only mostly... there's one mistake: the destination rect should be equal to the source rect, not the size of the destination canvas. I *did* have that right last time; gotta stop second-guessing. r=you with that change?

[WebKit Review Bot]

Comment on attachment 125419 [details]
patch

Attachment 125419 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/11424456

New failing tests:
fast/canvas/canvas-composite-canvas.html
canvas/philip/tests/2d.fillStyle.parse.current.removed.html

(In reply to comment #3)
> (In reply to comment #2)
> > (From update of attachment 125419 [details] [details])
> > Makes sense
> 
> Only mostly... there's one mistake: the destination rect should be equal to the source rect, not the size of the destination canvas. I *did* have that right last time; gotta stop second-guessing. r=you with that change?

I was going to say yes, but what's this failure the EWS bot is reporting?

[Tim Horton]

(In reply to comment #5)
> (In reply to comment #3)
> > (In reply to comment #2)
> > > (From update of attachment 125419 [details] [details] [details])
> > > Makes sense
> > 
> > Only mostly... there's one mistake: the destination rect should be equal to the source rect, not the size of the destination canvas. I *did* have that right last time; gotta stop second-guessing. r=you with that change?
> 
> I was going to say yes, but what's this failure the EWS bot is reporting?

I'm going to double check in a minute, but I'm willing to bet that *is* the failure the EWS bot is reporting. For some reason I didn't hit that when running the tests locally, though, so I have to figure out how to get it to fail!

Could you post the corrected version of the patch to see what the bit thinks about it?

[Tim Horton]

Created attachment 125463 [details]
revised patch

Sure, let's see what EWS says.

[Gyuyoung Kim]

Comment on attachment 125463 [details]
revised patch

Attachment 125463 [details] did not pass efl-ews (efl):
Output: http://queues.webkit.org/results/11421531

[Early Warning System Bot]

Comment on attachment 125463 [details]
revised patch

Attachment 125463 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/11416605

[Tim Horton]

Created attachment 125465 [details]
revised patch (which actually builds!)

Ugh, I undo'd a little too far.

[Tim Horton]

Landed in http://trac.webkit.org/changeset/106729; thanks, Dan!