[preload] Mandatory `as` value and related spec alignments

[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection of
  supported `as` values (instead of the `onerror` based fallback, which we're 
  deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered - 
  simplify HTML's processing model.

BUG=726647, 715690, 698520
Review-Url: https://codereview.chromium.org/2903653005
Cr-Commit-Position: refs/heads/master@{#476840}
Committed: https://chromium.googlesource.com/chromium/src/+/1f18863cc9b2a5fe463671f3eb60de2f95bbf283

[Yoav Weiss, 2017-05-24 09:30:33]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-24 09:31:10]

Dry run: CQ is trying da patch.

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

[Yoav Weiss, 2017-05-24 09:37:49]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-24 09:38:04]

Dry run: 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, 2017-05-24 11:04:55]

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

[commit-bot: I haz the power, 2017-05-24 11:04:56]

Dry run: Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Yoav Weiss, 2017-05-24 11:29:29]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-24 11:29:47]

Dry run: 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, 2017-05-24 12:45:42]

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

[commit-bot: I haz the power, 2017-05-24 12:45:42]

Dry run: Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Yoav Weiss, 2017-05-24 13:16:48]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-24 13:16:58]

Dry run: 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, 2017-05-24 14:44:14]

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

[commit-bot: I haz the power, 2017-05-24 14:44:15]

Dry run: Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[mizilaslan, 2017-05-24 22:58:40]

On 2017/05/24 14:44:15, commit-bot: I haz the power wrote:
> Dry run: Try jobs failed on following builders:
>   linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
>
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

Mehmet KIZILASLAN
Savcili Mah simsek sk
No 20 Kat 1.27020
Turkey/Gaziantep

[mizilaslan, 2017-05-24 22:59:04]

On 2017/05/24 22:58:40, mizilaslan wrote:
> On 2017/05/24 14:44:15, commit-bot: I haz the power wrote:
> > Dry run: Try jobs failed on following builders:
> >   linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
> >
>
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
> 
> Mehmet KIZILASLAN
> Savcili Mah simsek sk
> No 20 Kat 1.27020
> Turkey/Gaziantep

[Yoav Weiss, 2017-05-26 05:35:45]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-26 05:36:43]

Dry run: CQ is trying da patch.

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

[Yoav Weiss, 2017-05-26 05:36:52]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-26 05:38:00]

Dry run: 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, 2017-05-26 07:35:42]

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

[commit-bot: I haz the power, 2017-05-26 07:35:44]

Dry run: This issue passed the CQ dry run.

[Yoav Weiss, 2017-05-26 09:25:15]

Description was changed from

==========
Mandatory as value, Reflected as limited to known values, No error event for
unknown values

BUG=
==========

to

==========
Mandatory as value, Reflected as limited to known values, No error event for
unknown values

BUG=726647
==========

[Yoav Weiss, 2017-05-26 09:25:16]

yoav@yoav.ws changed reviewers:
- mizilaslan@gmail.com

[Yoav Weiss, 2017-05-26 09:27:36]

Description was changed from

==========
Mandatory as value, Reflected as limited to known values, No error event for
unknown values

BUG=726647
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where
developers omit the `as` value from their preload links, which often results in
double downloads. We've also had developer feedback that an empty `as` value
being mapped to XHR/fetch() feels magical. This change addresses that, by
defining a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of supported `as` values (instead of the `onerror` based fallback, which we're
deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- simplify HTML's processing model.

BUG=726647
==========

[Yoav Weiss, 2017-05-26 09:27:37]

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

[Yoav Weiss, 2017-05-26 09:44:23]

Description was changed from

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where
developers omit the `as` value from their preload links, which often results in
double downloads. We've also had developer feedback that an empty `as` value
being mapped to XHR/fetch() feels magical. This change addresses that, by
defining a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of supported `as` values (instead of the `onerror` based fallback, which we're
deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- simplify HTML's processing model.

BUG=726647
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where
developers omit the `as` value from their preload links, which often results in
double downloads. We've also had developer feedback that an empty `as` value
being mapped to XHR/fetch() feels magical. This change addresses that, by
defining a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of supported `as` values (instead of the `onerror` based fallback, which we're
deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- simplify HTML's processing model.

BUG=726647, 715690
==========

[Mike West, 2017-05-28 06:21:09]

Implementation looks good! I have some questions about the tests, though.

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
File
third_party/WebKit/LayoutTests/external/wpt/html/dom/reflection-metadata-expected.txt
(right):

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/external/wpt/html/dom/reflection-metadata-expected.txt:37:
FAIL link.as: setAttribute() to "DOCUMENT" assert_equals: IDL get expected
"document" but got ""
Are these failures bugs in the test, or bugs in our implementation? Skimming the
spec, they look like bugs in the test, in which case we should update the test.
Is that accurate?

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
File third_party/WebKit/LayoutTests/http/tests/preload/onerror_event.html
(left):

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/http/tests/preload/onerror_event.html:1:
<!DOCTYPE html>
Why drop this test? Is it replicated elsewhere?

[Charlie Harrison, 2017-05-28 22:22:43]

I think this change is nice but I'm wondering if we're far enough along in the
blink-dev thread to move forward in this patch. E.g. we might need a deprecation
for one milestone

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/HTMLLinkElement.idl (right):

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/HTMLLinkElement.idl:35: [Reflect,
ReflectOnly=("","script","style","image","video", "audio", "track", "font",
"fetch"), ReflectMissing="", ReflectInvalid=""] attribute DOMString as;
Not an expert on idl, does this enforce case sensitivity?

[Yoav Weiss, 2017-05-31 05:00:41]

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
File
third_party/WebKit/LayoutTests/external/wpt/html/dom/reflection-metadata-expected.txt
(right):

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/external/wpt/html/dom/reflection-metadata-expected.txt:37:
FAIL link.as: setAttribute() to "DOCUMENT" assert_equals: IDL get expected
"document" but got ""
On 2017/05/28 06:21:08, Mike West wrote:
> Are these failures bugs in the test, or bugs in our implementation? Skimming
the
> spec, they look like bugs in the test, in which case we should update the
test.
> Is that accurate?

These failures are resulting from the fact that we don't (yet) support
preloading for all these destinations, so they are expected.

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
File third_party/WebKit/LayoutTests/http/tests/preload/onerror_event.html
(left):

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/http/tests/preload/onerror_event.html:1:
<!DOCTYPE html>
On 2017/05/28 06:21:08, Mike West wrote:
> Why drop this test? Is it replicated elsewhere?

/external/wpt/preload/onerror_event.html is an identical test

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/HTMLLinkElement.idl (right):

https://codereview.chromium.org/2903653005/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/HTMLLinkElement.idl:35: [Reflect,
ReflectOnly=("","script","style","image","video", "audio", "track", "font",
"fetch"), ReflectMissing="", ReflectInvalid=""] attribute DOMString as;
On 2017/05/28 22:22:43, Charlie Harrison wrote:
> Not an expert on idl, does this enforce case sensitivity?

It enforces case-sensiticity on the output, but not on the input. e.g. an `as`
value for "VidEo" will be reflected as "video".
See
https://chromium.googlesource.com/chromium/src/+/master/third_party/WebKit/So...

I'll add such values to the reflected_as_attribute test

[Yoav Weiss, 2017-05-31 05:02:18]

On 2017/05/28 22:22:43, Charlie Harrison wrote:
> I think this change is nice but I'm wondering if we're far enough along in the
> blink-dev thread to move forward in this patch. E.g. we might need a
deprecation
> for one milestone
> 

I'm not planning to land this until the I2S has 3 LGTMs

[Yoav Weiss, 2017-05-31 05:08:34]

The CQ bit was checked by yoav@yoav.ws to run a CQ dry run

[commit-bot: I haz the power, 2017-05-31 05:08:51]

Dry run: 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, 2017-05-31 07:26:53]

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

[commit-bot: I haz the power, 2017-05-31 07:26:54]

Dry run: This issue passed the CQ dry run.

[Yoav Weiss, 2017-05-31 07:29:59]

Description was changed from

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where
developers omit the `as` value from their preload links, which often results in
double downloads. We've also had developer feedback that an empty `as` value
being mapped to XHR/fetch() feels magical. This change addresses that, by
defining a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of supported `as` values (instead of the `onerror` based fallback, which we're
deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- simplify HTML's processing model.

BUG=726647, 715690
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where
developers omit the `as` value from their preload links, which often results in
double downloads. We've also had developer feedback that an empty `as` value
being mapped to XHR/fetch() feels magical. This change addresses that, by
defining a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of supported `as` values (instead of the `onerror` based fallback, which we're
deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

[Charlie Harrison, 2017-05-31 12:26:01]

On 2017/05/31 05:02:18, Yoav Weiss wrote:
> On 2017/05/28 22:22:43, Charlie Harrison wrote:
> > I think this change is nice but I'm wondering if we're far enough along in
the
> > blink-dev thread to move forward in this patch. E.g. we might need a
> deprecation
> > for one milestone
> > 
> 
> I'm not planning to land this until the I2S has 3 LGTMs

Gotcha I was more thinking if the response from i2s was "we should do a
deprecation for 1 milestone" in which case this CL would be on hold for a while.
It looks like it's probably not going that way now though.

Let me do one more once over of all the tests... :)

[Charlie Harrison, 2017-05-31 12:36:51]

LGTM but would you wrap the CL desc to 80 cols?

I will be happy never seeing a link preload with an empty "as" tag again :)

[Yoav Weiss, 2017-05-31 12:59:28]

Description was changed from

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where
developers omit the `as` value from their preload links, which often results in
double downloads. We've also had developer feedback that an empty `as` value
being mapped to XHR/fetch() feels magical. This change addresses that, by
defining a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of supported `as` values (instead of the `onerror` based fallback, which we're
deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
*Properly reflect a limited set of known `as` values - enable feature detection
of
  supported `as` values (instead of the `onerror` based fallback, which we're 
  deprecating.
*Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
  simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

[Yoav Weiss, 2017-05-31 12:59:56]

Description was changed from

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
*Properly reflect a limited set of known `as` values - enable feature detection
of
  supported `as` values (instead of the `onerror` based fallback, which we're 
  deprecating.
*Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
  simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of
   supported `as` values (instead of the `onerror` based fallback, which we're 
   deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
   simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

[Yoav Weiss, 2017-05-31 13:00:15]

Description was changed from

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of
   supported `as` values (instead of the `onerror` based fallback, which we're 
   deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
   simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of
  supported `as` values (instead of the `onerror` based fallback, which we're 
  deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
  simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

[Yoav Weiss, 2017-05-31 13:01:10]

On 2017/05/31 12:36:51, Charlie Harrison wrote:
> LGTM but would you wrap the CL desc to 80 cols?

Wrapped! Thanks for reviewing!!

> 
> I will be happy never seeing a link preload with an empty "as" tag again :)

Me too :D

[domenic, 2017-06-02 21:27:45]

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

[domenic, 2017-06-02 21:27:46]

https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Lay...
File third_party/WebKit/LayoutTests/external/wpt/preload/reflected-as-value.html
(right):

https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/external/wpt/preload/reflected-as-value.html:1:
<!DOCTYPE html>
I think this test is redundant with the mega-reflection test, but I guess it's
fine to have extras.

https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Sou...
File third_party/WebKit/Source/core/html/HTMLLinkElement.idl (right):

https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Sou...
third_party/WebKit/Source/core/html/HTMLLinkElement.idl:35: [Reflect,
ReflectOnly=("","script","style","image","video", "audio", "track", "font",
"fetch"), ReflectMissing="", ReflectInvalid=""] attribute DOMString as;
I think "" should not be included in the list. You may also not need the
ReflectMissing/ReflectInvalid.

[Yoav Weiss, 2017-06-02 21:34:16]

The CQ bit was checked by yoav@yoav.ws

[Yoav Weiss, 2017-06-02 21:34:35]

The CQ bit was unchecked by yoav@yoav.ws

[Yoav Weiss, 2017-06-02 21:47:10]

The CQ bit was checked by yoav@yoav.ws

[Yoav Weiss, 2017-06-02 21:47:10]

The patchset sent to the CQ was uploaded after l-g-t-m from
csharrison@chromium.org
Link to the patchset: https://codereview.chromium.org/2903653005/#ps140001
(title: "IDL changes")

[commit-bot: I haz the power, 2017-06-02 21:47:54]

CQ is trying da patch.

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

[Yoav Weiss, 2017-06-02 21:49:51]

On 2017/06/02 21:27:46, domenic wrote:
>
https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Lay...
> File
third_party/WebKit/LayoutTests/external/wpt/preload/reflected-as-value.html
> (right):
> 
>
https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Lay...
> third_party/WebKit/LayoutTests/external/wpt/preload/reflected-as-value.html:1:
> <!DOCTYPE html>
> I think this test is redundant with the mega-reflection test, but I guess it's
> fine to have extras.

Yeah, I preferred to have an extra test explicitly for `as`

> 
>
https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Sou...
> File third_party/WebKit/Source/core/html/HTMLLinkElement.idl (right):
> 
>
https://codereview.chromium.org/2903653005/diff/120001/third_party/WebKit/Sou...
> third_party/WebKit/Source/core/html/HTMLLinkElement.idl:35: [Reflect,
> ReflectOnly=("","script","style","image","video", "audio", "track", "font",
> "fetch"), ReflectMissing="", ReflectInvalid=""] attribute DOMString as;
> I think "" should not be included in the list. You may also not need the
> ReflectMissing/ReflectInvalid.

Removed

[commit-bot: I haz the power, 2017-06-03 00:20:48]

CQ is committing da patch.

Bot data: {"patchset_id": 140001, "attempt_start_ts": 1496440030169940,
"parent_rev": "37003a1722d89acc689cd3911efceb583bad43e0", "commit_rev":
"1f18863cc9b2a5fe463671f3eb60de2f95bbf283"}

[commit-bot: I haz the power, 2017-06-03 00:20:59]

Description was changed from

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of
  supported `as` values (instead of the `onerror` based fallback, which we're 
  deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
  simplify HTML's processing model.

BUG=726647, 715690, 698520
==========

to

==========
[preload] Mandatory `as` value and related spec alignments

The motivation for this change is to avoid double downloads following developer
confusion around empty `as` values, enable feature detection of supported `as`
values, and simplify the HTML processing model. Details below:
* Avoid fetching a resource if `as` is missing - we've seen many cases where 
  developers omit the `as` value from their preload links, which often results
in 
  double downloads. We've also had developer feedback that an empty `as` value
  being mapped to XHR/fetch() feels magical. This change addresses that, by
defining
  a "fetch" value for these cases and ignoring empty/missing `as` values.
* Properly reflect a limited set of known `as` values - enable feature detection
of
  supported `as` values (instead of the `onerror` based fallback, which we're 
  deprecating.
* Avoid triggering the "onerror" event when an invalid `as` value is encountered
- 
  simplify HTML's processing model.

BUG=726647, 715690, 698520

Review-Url: https://codereview.chromium.org/2903653005
Cr-Commit-Position: refs/heads/master@{#476840}
Committed:
https://chromium.googlesource.com/chromium/src/+/1f18863cc9b2a5fe463671f3eb60...
==========

[commit-bot: I haz the power, 2017-06-03 00:21:02]

Committed patchset #8 (id:140001) as
https://chromium.googlesource.com/chromium/src/+/1f18863cc9b2a5fe463671f3eb60...