predictors: Add Manifest table to ResourcePrefetchPredictor.

predictors: Add Manifest table to ResourcePrefetchPredictor.

The very first step to reuse Bork manifests in ResourcePrefetchPredictor. This
CL adds a new table to the predictor database.

BUG=699115
Review-Url: https://codereview.chromium.org/2738613003
Cr-Commit-Position: refs/heads/master@{#455798}
Committed: https://chromium.googlesource.com/chromium/src/+/8b3c17d1e53fade1111ec0fe9aed8a6f891d5331

[alexilin, 2017-03-07 16:08:14]

alexilin@chromium.org changed reviewers:
+ lizeb@chromium.org

[alexilin, 2017-03-07 16:08:16]

Hi Benoit,

This is a first CL to bring power of precache to ResourcePrefetchPredictor.
PTAL!

https://codereview.chromium.org/2738613003/diff/1/chrome/browser/predictors/r...
File chrome/browser/predictors/resource_prefetch_predictor_tables.h (left):

https://codereview.chromium.org/2738613003/diff/1/chrome/browser/predictors/r...
chrome/browser/predictors/resource_prefetch_predictor_tables.h:40: // Map from
primary key to PrefetchData for the key.
Get rid of obvious comments.

https://codereview.chromium.org/2738613003/diff/1/chrome/browser/predictors/r...
File chrome/browser/predictors/resource_prefetch_predictor_test_util.h (right):

https://codereview.chromium.org/2738613003/diff/1/chrome/browser/predictors/r...
chrome/browser/predictors/resource_prefetch_predictor_test_util.h:87: namespace
precache {
I know, it doesn't look good but it won't compile if put operator== overload
into predictors namespace. We had the same problem earlier with this file.

[alexilin, 2017-03-08 09:10:16]

I've just realized that we don't have a date field in a manifest that's
unfortunate.
I'm tending to define our own proto and use it instead of
precache::PrecacheManifest.

[Benoit L, 2017-03-08 15:03:09]

Thanks!

One high-level question: How do we make sure that the database version gets
updated when the protocol buffer does?

[alexilin, 2017-03-08 15:22:00]

On 2017/03/08 15:03:09, Benoit L wrote:
> Thanks!
> 
> One high-level question: How do we make sure that the database version gets
> updated when the protocol buffer does?

That's a very good question. This is another sign to define our own 
proto. I think, I'll just implement it.
Alternatively, we could add a comment to precache.proto saying that they need 
to update a version in a database that they don't depend on.

[alexilin, 2017-03-09 16:40:08]

Caution comment was added.
CL is ready to be reviewed.

[Benoit L, 2017-03-09 17:00:28]

lgtm, thanks.

https://codereview.chromium.org/2738613003/diff/1/chrome/browser/predictors/r...
File chrome/browser/predictors/resource_prefetch_predictor_test_util.h (right):

https://codereview.chromium.org/2738613003/diff/1/chrome/browser/predictors/r...
chrome/browser/predictors/resource_prefetch_predictor_test_util.h:87: namespace
precache {
On 2017/03/07 16:08:15, alexilin wrote:
> I know, it doesn't look good but it won't compile if put operator== overload
> into predictors namespace. We had the same problem earlier with this file.


Yes, I remember.
Sigh...

[alexilin, 2017-03-09 17:07:56]

The CQ bit was checked by alexilin@chromium.org

[commit-bot: I haz the power, 2017-03-09 17:08:38]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-03-09 18:27:46]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1489079276410830,
"parent_rev": "55462c0bffc90caf9c5e11bdd6a5d611f187330d", "commit_rev":
"8b3c17d1e53fade1111ec0fe9aed8a6f891d5331"}

[commit-bot: I haz the power, 2017-03-09 18:28:31]

Description was changed from

==========
predictors: Add Manifest table to ResourcePrefetchPredictor.

The very first step to reuse Bork manifests in ResourcePrefetchPredictor. This
CL adds a new table to the predictor database.

BUG=699115
==========

to

==========
predictors: Add Manifest table to ResourcePrefetchPredictor.

The very first step to reuse Bork manifests in ResourcePrefetchPredictor. This
CL adds a new table to the predictor database.

BUG=699115

Review-Url: https://codereview.chromium.org/2738613003
Cr-Commit-Position: refs/heads/master@{#455798}
Committed:
https://chromium.googlesource.com/chromium/src/+/8b3c17d1e53fade1111ec0fe9aed...
==========

[commit-bot: I haz the power, 2017-03-09 18:28:32]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/chromium/src/+/8b3c17d1e53fade1111ec0fe9aed...