Add HTMLPreloadScanner support for <link rel=preload>

Add HTMLPreloadScanner support for <link rel=preload>

HTMLPreloadScanner should support <link rel=preload> so that these tags
would be discovered as soon as possible, even if the parser is blocked.
This CL adds that support.

BUG=552290
Committed: https://crrev.com/17a8c60eb46b6617f68d8c450988d15a18cb88dc
Cr-Commit-Position: refs/heads/master@{#368826}

[Yoav Weiss, 2016-01-07 15:12:33]

yoav@yoav.ws changed reviewers:
+ mkwst@chromium.org

[Yoav Weiss, 2016-01-07 15:12:33]

Hey Mike,

Can you take a look? It's a fairly small and simple one :)

[Yoav Weiss, 2016-01-07 16:03:35]

On 2016/01/07 15:12:33, Yoav Weiss wrote:
> Hey Mike,
> 
> Can you take a look? It's a fairly small and simple one :)

test failures are related. on it

[Yoav Weiss, 2016-01-08 10:11:17]

yoav@yoav.ws changed reviewers:
+ japhet@chromium.org

[Yoav Weiss, 2016-01-08 10:11:17]

Fixed tests and bugs they found (hopefully. We'll see what the bots say).

japhet - can you take a look? Mike's status says he is out and about.

[Nate Chapin, 2016-01-12 03:28:17]

LGTM

https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/html/parser/PreloadRequest.cpp (right):

https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/html/parser/PreloadRequest.cpp:52:
request.setAvoidBlockingOnLoad(true);
It's unfortunate to need to set this in a second place, but I don't see a good
way to avoid that.

[Yoav Weiss, 2016-01-12 06:37:57]

https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/html/parser/PreloadRequest.cpp (right):

https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/html/parser/PreloadRequest.cpp:52:
request.setAvoidBlockingOnLoad(true);
On 2016/01/12 03:28:17, Nate Chapin wrote:
> It's unfortunate to need to set this in a second place, but I don't see a good
> way to avoid that.

agree

[Yoav Weiss, 2016-01-12 06:38:05]

The CQ bit was checked by yoav@yoav.ws

[commit-bot: I haz the power, 2016-01-12 06:39:01]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1563263002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1563263002/20001

[commit-bot: I haz the power, 2016-01-12 08:43:27]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-01-12 08:44:30]

Description was changed from

==========
Add HTMLPreloadScanner support for <link rel=preload>

HTMLPreloadScanner should support <link rel=preload> so that these tags
would be discovered as soon as possible, even if the parser is blocked.
This CL adds that support.

BUG=552290
==========

to

==========
Add HTMLPreloadScanner support for <link rel=preload>

HTMLPreloadScanner should support <link rel=preload> so that these tags
would be discovered as soon as possible, even if the parser is blocked.
This CL adds that support.

BUG=552290

Committed: https://crrev.com/17a8c60eb46b6617f68d8c450988d15a18cb88dc
Cr-Commit-Position: refs/heads/master@{#368826}
==========

[commit-bot: I haz the power, 2016-01-12 08:44:31]

Patchset 2 (id:??) landed as
https://crrev.com/17a8c60eb46b6617f68d8c450988d15a18cb88dc
Cr-Commit-Position: refs/heads/master@{#368826}

[igrigorik, 2016-01-12 22:11:54]

igrigorik@chromium.org changed reviewers:
+ igrigorik@chromium.org

[igrigorik, 2016-01-12 22:11:55]

https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
File third_party/WebKit/Source/core/loader/LinkLoader.cpp (right):

https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
third_party/WebKit/Source/core/loader/LinkLoader.cpp:183: // TODO(yoav): Is this
correct? If as is missing or invalid, it should be subject to "connect-src" CSP
directives.
(sorry for late review). To answer the question: yes, I think connect-src is the
right default.. we describe preload s declarative fetch(), so it makes sense to
default to connect-src. 

That said, does LinkSubresource default to this policy?

[Yoav Weiss, 2016-01-13 07:07:52]

On 2016/01/12 22:11:55, igrigorik wrote:
>
https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
> File third_party/WebKit/Source/core/loader/LinkLoader.cpp (right):
> 
>
https://codereview.chromium.org/1563263002/diff/20001/third_party/WebKit/Sour...
> third_party/WebKit/Source/core/loader/LinkLoader.cpp:183: // TODO(yoav): Is
this
> correct? If as is missing or invalid, it should be subject to "connect-src"
CSP
> directives.
> (sorry for late review). To answer the question: yes, I think connect-src is
the
> right default.. we describe preload s declarative fetch(), so it makes sense
to
> default to connect-src. 
> 
> That said, does LinkSubresource default to this policy?

That's exactly why this TODO is here :) I need to look into it and fix it before
shipping.