cc: Reset raster scale if we didn't sync high res tiling.

cc: Reset raster scale if we didn't sync high res tiling.

When we sync tilings, we always expect to sync the high res
tiling. However, in situations where our minimum contents scale
prevents us from syncing the high res tiling, we should instead
reset the raster contents scale, since those would be invalid
as well. This also ensures that we will create a new high res
tiling when we call ManageTilings.

BUG=374143
R=enne

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=271527

[vmpstr, 2014-05-16 22:56:04]

Please take a look. It's possible that maybe DrawsContent should return false if
we have empty bounds or something, but then we won't know to set the tiling set
bounds to empty? This seems like a more conservative approach that also covers
cases where the twin might have all tilings less than minimum contents scale (is
that possible?)

[enne (OOO), 2014-05-19 17:24:07]

How does this prevent the crash in the bug?

[vmpstr, 2014-05-19 19:31:18]

On 2014/05/19 17:24:07, enne wrote:
> How does this prevent the crash in the bug?

This just prevents a DCHECK in ManageTilings that ensures that if we have 0
tilings, we don't have a raster page scale. Please hold off on the review, as I
think there's more work to be done here.

[vmpstr, 2014-05-19 21:57:54]

PTAL

[enne (OOO), 2014-05-19 22:27:36]

https://codereview.chromium.org/294463002/diff/40001/cc/layers/picture_layer_...
File cc/layers/picture_layer_impl_unittest.cc (right):

https://codereview.chromium.org/294463002/diff/40001/cc/layers/picture_layer_...
cc/layers/picture_layer_impl_unittest.cc:1756: // Ensure we need post commit
initialization on the pending layer.
This is a very informative comment.  Maybe say you're simulating a commit where
this layer has changed its bounds? Normally the "post-commit init" would happen
during commit because of this change, but if you're faking the commit, setting
this flag will make it happen during UDP instead (I think?).

[vmpstr, 2014-05-19 22:46:28]

PTAL

https://codereview.chromium.org/294463002/diff/40001/cc/layers/picture_layer_...
File cc/layers/picture_layer_impl_unittest.cc (right):

https://codereview.chromium.org/294463002/diff/40001/cc/layers/picture_layer_...
cc/layers/picture_layer_impl_unittest.cc:1756: // Ensure we need post commit
initialization on the pending layer.
On 2014/05/19 22:27:36, enne wrote:
> This is a very informative comment.  Maybe say you're simulating a commit
where
> this layer has changed its bounds? Normally the "post-commit init" would
happen
> during commit because of this change, but if you're faking the commit, setting
> this flag will make it happen during UDP instead (I think?).

Done.

[enne (OOO), 2014-05-19 22:53:42]

lgtm, thanks!

[vmpstr, 2014-05-19 22:53:57]

The CQ bit was checked by vmpstr@chromium.org

[commit-bot: I haz the power, 2014-05-19 22:55:08]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/vmpstr@chromium.org/294463002/60001

[commit-bot: I haz the power, 2014-05-20 01:09:23]

Change committed as 271527