Flush the cairo surface before we try to write to it.

Flush the cairo surface before we try to write to it.

Newer versions of Cairo assert if the surface isn't flushed before it is marked dirty but this should be done anyway to ensure correct layering of draw operations.

BUG=65692
TEST=No crash on Maverick (Ubuntu 10.10)

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=69334

[vandebo (ex-Chrome), 2010-12-14 22:22:14]

Evan, do you have time to look at this?

[Evan Martin, 2010-12-14 22:27:11]

LGTM

+estade who dealt with some of this in the past

http://codereview.chromium.org/5737004/diff/1/skia/ext/bitmap_platform_device...
File skia/ext/bitmap_platform_device_linux.cc (right):

http://codereview.chromium.org/5737004/diff/1/skia/ext/bitmap_platform_device...
skia/ext/bitmap_platform_device_linux.cc:158: cairo_surface_flush(surface);
Doesn't this belong in endPlatformPaint()?
It seems from the docs that you're supposed to call it before using non-cairo
APIs to work on the surface... I guess that's what the mark_dirty is for,
though, huh.

[vandebo (ex-Chrome), 2010-12-14 22:47:53]

Thanks for looking.

http://codereview.chromium.org/5737004/diff/1/skia/ext/bitmap_platform_device...
File skia/ext/bitmap_platform_device_linux.cc (right):

http://codereview.chromium.org/5737004/diff/1/skia/ext/bitmap_platform_device...
skia/ext/bitmap_platform_device_linux.cc:158: cairo_surface_flush(surface);
On 2010/12/14 22:27:11, Evan Martin wrote:
> Doesn't this belong in endPlatformPaint()?
> It seems from the docs that you're supposed to call it before using non-cairo
> APIs to work on the surface... I guess that's what the mark_dirty is for,
> though, huh.

I read the Docs the other way around.  flush needs to be called before you
modify the surface directly (cairo may have things it hasn't finalized to the
surface).  Then, if you modify the surface, you need to call mark_dirty before
making further cairo calls so that cairo knows to invalidate any cached state it
has.  Asserts in new versions of mark_dirty reinforce this interpretation.

Linux doesn't currently have an endPlatformPaint.  I assume this is ok and
currently works because mark_dirty just flushes cairo caches and doesn't
recompute anything.

[vandebo (ex-Chrome), 2010-12-14 22:47:53]

Thanks for looking.