cc: Make eviction queue building consistent with raster.

cc: Make eviction queue building consistent with raster.

This patch changes the way eviction queue is built to be more like the
raster queue. That is, instead of taking a pointer to the queue to be
built, the function returns a scoped_ptr to the queue that it built.

R=reveman, danakj

Committed: https://crrev.com/997b00ea9a6a101266c32ce26806c7584940d12e
Cr-Commit-Position: refs/heads/master@{#312925}

[vmpstr, 2015-01-22 20:02:22]

Please take a look. This is to make it consistent with a patch in
https://codereview.chromium.org/863013004/

[danakj, 2015-01-22 23:12:23]

https://codereview.chromium.org/862263004/diff/20001/cc/resources/tile_manage...
File cc/resources/tile_manager.cc (right):

https://codereview.chromium.org/862263004/diff/20001/cc/resources/tile_manage...
cc/resources/tile_manager.cc:643: eviction_priority_queue_.reset();
Can you explain this? There's a DCHECK above that it's null. What in between
will make a queue?

[vmpstr, 2015-01-22 23:45:03]

https://codereview.chromium.org/862263004/diff/20001/cc/resources/tile_manage...
File cc/resources/tile_manager.cc (right):

https://codereview.chromium.org/862263004/diff/20001/cc/resources/tile_manage...
cc/resources/tile_manager.cc:643: eviction_priority_queue_.reset();
On 2015/01/22 23:12:23, danakj wrote:
> Can you explain this? There's a DCHECK above that it's null. What in between
> will make a queue?

Since creating an eviction queue might be expensive, we only create it if we
need it in FreeTileResources* function call. Since that eviction queue becomes
stale at the end of this call, I reset it so in the future we know to create a
new one (and that's what the DCHECK is checking). I've added a comment similar
to this. Please see if it's more clear.

I could also basically replace each call of
"eviction_priority_queue_is_up_to_date_" to eviction_priority_queue_.reset() if
that's cleaner. I kind of decided to do it after, instead of before the loop but
it shouldn't matter in practice.

[vmpstr, 2015-01-23 00:19:41]

PTAL. After some discussion I changed the eviction queue in tile manager to be a
local variable.

[danakj, 2015-01-23 00:24:41]

LGTM

[vmpstr, 2015-01-23 18:32:51]

The CQ bit was checked by vmpstr@chromium.org

[commit-bot: I haz the power, 2015-01-23 18:33:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/862263004/80001

[vmpstr, 2015-01-23 18:52:48]

Patchset #7 (id:120001) has been deleted

[vmpstr, 2015-01-23 18:52:55]

Patchset #6 (id:100001) has been deleted

[vmpstr, 2015-01-23 18:52:59]

The CQ bit was checked by vmpstr@chromium.org

[commit-bot: I haz the power, 2015-01-23 18:54:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/862263004/80001

[commit-bot: I haz the power, 2015-01-23 19:53:57]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-01-23 19:53:58]

Try jobs failed on following builders:
  mac_chromium_rel_ng on tryserver.chromium.mac
(http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[vmpstr, 2015-01-23 21:01:28]

The CQ bit was checked by vmpstr@chromium.org

[commit-bot: I haz the power, 2015-01-23 21:02:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/862263004/80001

[commit-bot: I haz the power, 2015-01-23 21:03:35]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-01-23 21:04:39]

Patchset 5 (id:??) landed as
https://crrev.com/997b00ea9a6a101266c32ce26806c7584940d12e
Cr-Commit-Position: refs/heads/master@{#312925}