Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no
longer present in the standard. Regardless, the overloads should
never execute in Service Workers as those types (MediaStream,
MediaSource) are not exposed in that context.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread: https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoeyBwAJ

BUG=604951, 608460
Committed: https://crrev.com/9e232c512c962a7aecc85e07f7a48a4735dc7a25
Cr-Commit-Position: refs/heads/master@{#393271}

[jsbell, 2016-04-21 23:55:17]

jsbell@chromium.org changed reviewers:
+ haraken@chromium.org

[jsbell, 2016-04-21 23:55:18]

haraken@, can you do an initial review for the bindings change and/or suggest an
alternate reviewer?

(I'll still need an API_OWNERS review as this technically unships an API, but
it's scoped to SW and would never have worked usefully.)

[haraken, 2016-04-22 00:11:00]

LGTM

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/m...
File third_party/WebKit/Source/modules/mediasource/URLMediaSource.idl (right):

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/m...
third_party/WebKit/Source/modules/mediasource/URLMediaSource.idl:34:
[Exposed=(Window,DedicatedWorker,SharedWorker), CallWith=ExecutionContext]
static DOMString? createObjectURL(MediaSource source);

You're changing the exposed policy from [Exposed=(Window,Worker)] on URL.idl to
[Exposed=(Window,DedicatedWorker,SharedWorker)]. This is expected, right?

[jsbell, 2016-04-22 00:16:26]

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/m...
File third_party/WebKit/Source/modules/mediasource/URLMediaSource.idl (right):

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/m...
third_party/WebKit/Source/modules/mediasource/URLMediaSource.idl:34:
[Exposed=(Window,DedicatedWorker,SharedWorker), CallWith=ExecutionContext]
static DOMString? createObjectURL(MediaSource source);
On 2016/04/22 00:10:59, haraken wrote:
> 
> You're changing the exposed policy from [Exposed=(Window,Worker)] on URL.idl
to
> [Exposed=(Window,DedicatedWorker,SharedWorker)]. This is expected, right?

Yes, just for the partial interface defined in File API, though. This
corresponds to this spec change:

https://github.com/w3c/FileAPI/commit/b9c2275df53cf3a808f1f272f3d6134d1b920549

This should explicitly exclude ServiceWorker and future things like
CompositorWorker.

The partials for MediaSource and MediaStream no longer exist in specs, but the
same logic should apply - the minting URLs in SWs doesn't make sense. Should
loop in OWNERS for those modules as a sanity check?

[haraken, 2016-04-22 00:30:59]

On 2016/04/22 00:16:26, jsbell wrote:
>
https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/m...
> File third_party/WebKit/Source/modules/mediasource/URLMediaSource.idl (right):
> 
>
https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/m...
> third_party/WebKit/Source/modules/mediasource/URLMediaSource.idl:34:
> [Exposed=(Window,DedicatedWorker,SharedWorker), CallWith=ExecutionContext]
> static DOMString? createObjectURL(MediaSource source);
> On 2016/04/22 00:10:59, haraken wrote:
> > 
> > You're changing the exposed policy from [Exposed=(Window,Worker)] on URL.idl
> to
> > [Exposed=(Window,DedicatedWorker,SharedWorker)]. This is expected, right?
> 
> Yes, just for the partial interface defined in File API, though. This
> corresponds to this spec change:
> 
> https://github.com/w3c/FileAPI/commit/b9c2275df53cf3a808f1f272f3d6134d1b920549
> 
> This should explicitly exclude ServiceWorker and future things like
> CompositorWorker.
> 
> The partials for MediaSource and MediaStream no longer exist in specs, but the
> same logic should apply - the minting URLs in SWs doesn't make sense. Should
> loop in OWNERS for those modules as a sanity check?

Thanks for the clarification! Makes sense.

[sof, 2016-04-22 06:19:28]

sigbjornf@opera.com changed reviewers:
+ sigbjornf@opera.com

[sof, 2016-04-22 06:19:29]

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/URLFileAPI.cpp (right):

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/URLFileAPI.cpp:19: if (!executionContext)
assert instead? i.e., can the current execution context being passed in ever be
null?

[jsbell, 2016-04-22 18:10:47]

On 2016/04/22 00:16:26, jsbell wrote:
> The partials for MediaSource and MediaStream no longer exist in specs, but the
> same logic should apply - the minting URLs in SWs doesn't make sense. Should
> loop in OWNERS for those modules as a sanity check?

I was incorrect about the state of MediaSource and filed a spec issue:

https://github.com/w3c/media-source/issues/67

[jsbell, 2016-04-22 18:14:34]

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/fileapi/URLFileAPI.cpp (right):

https://codereview.chromium.org/1908263002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/fileapi/URLFileAPI.cpp:19: if (!executionContext)
On 2016/04/22 06:19:29, sof wrote:
> assert instead? i.e., can the current execution context being passed in ever
be
> null?

That's how I found the code. :( I'll switch it to a DCHECK and see if it sticks.

[jsbell, 2016-04-22 18:19:11]

jsbell@chromium.org changed reviewers:
+ tommi@chromium.org, wolenetz@chromium.org

[jsbell, 2016-04-22 18:19:11]

wolenetz@ - can you review the modules/mediasource change?

tommi@ - can you review the modules/mediastream change?

See previous comments on the CL for a little more context.

[jsbell, 2016-04-22 18:21:14]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?) Also, a bindings template change was needed to
ensure the method supporting the partial interface / static method /
conditional exposure combination was exported for use in modules.

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688

BUG=604951
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?) Also, a bindings template change was needed to
ensure the method supporting the partial interface / static method /
conditional exposure combination was exported for use in modules.

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

BUG=604951
==========

[tommi (sloooow) - chröme, 2016-04-25 10:20:26]

lgtm

[jsbell, 2016-04-25 20:43:49]

FYI, I'm splitting the URLFileAPI-refactor part of this CL out on its own:

https://codereview.chromium.org/1921063002

I'll rebase this one to be just the [Exposed] changes.

[jsbell, 2016-04-25 20:47:48]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?) Also, a bindings template change was needed to
ensure the method supporting the partial interface / static method /
conditional exposure combination was exported for use in modules.

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

BUG=604951
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

BUG=604951
==========

[wolenetz, 2016-05-02 19:22:03]

lgtm (mediasource)
jsbell@ : is there a crbug tracking the createFor() in URLFileAPI.idl?
MediaSource might need similar (though currently
(https://github.com/w3c/media-source/issues/10#issuecomment-201024017) it looks
like lack of createFor() won't block MSE spec V1 reaching PR.

[jsbell, 2016-05-02 20:24:41]

On 2016/05/02 19:22:03, wolenetz wrote:
> lgtm (mediasource)
> jsbell@ : is there a crbug tracking the createFor() in URLFileAPI.idl?

Filed one: https://bugs.chromium.org/p/chromium/issues/detail?id=608460

[jsbell, 2016-05-02 20:25:00]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

BUG=604951
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

BUG=604951,608460
==========

[wolenetz, 2016-05-02 20:27:26]

On 2016/05/02 20:24:41, jsbell wrote:
> On 2016/05/02 19:22:03, wolenetz wrote:
> > lgtm (mediasource)
> > jsbell@ : is there a crbug tracking the createFor() in URLFileAPI.idl?
> 
> Filed one: https://bugs.chromium.org/p/chromium/issues/detail?id=608460

Thank you

[jsbell, 2016-05-02 21:08:45]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

BUG=604951,608460
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460
==========

[jsbell, 2016-05-09 23:03:57]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard. Regardless, the overloads should
never execute in Service Workers as those types (MediaStream,
MediaSource) are not exposed in that context.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460
==========

[jsbell, 2016-05-10 19:00:55]

jsbell@chromium.org changed reviewers:
+ yhirano@chromium.org

[jsbell, 2016-05-10 19:00:55]

yhirano@ - can you review the fetch test changes?

This makes thorough/scheme-blob*.html a no-op in serviceworker contexts. The
obvious alternatives are to not auto-generate thorough/scheme-blob or to teach
the generator to skip the serviceworker versions, which seem worse, but I'm
flexible.

[yhirano, 2016-05-11 01:54:05]

lgtm

Fetch tests are written in the service worker layout tests style[1]. Please keep
the style.

1: https://www.chromium.org/blink/serviceworker/testing

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js
(right):

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:77: });
+indent

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:78: },
'fetch blob: URL');
+indent

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:84:
t.unreached_func('fetching non-existent blob: URL must fail'),
-indent

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:85:
function() {});
-indent

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:86: },
'fetch non-existent blob: URL');
+indent

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/thorough/scheme-blob.js
(right):

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/thorough/scheme-blob.js:6:
// Only [Exposed=(Window,DedicatedWorker,SharedWorker)].
Could you define an empty TEST_TARGETS here and move
|executeTests(TEST_TARGETS)| back to the last?

[jsbell, 2016-05-11 21:52:40]

jsbell@chromium.org changed reviewers:
+ rbyers@chromium.org

[jsbell, 2016-05-11 21:52:42]

rbyers@ - can you API_OWNERS review?

yhirano@ - thanks for the detailed formatting feedback! Did we ever figure out
clang-format settings for that style? (since I suspect I'll never get emacs to
do it correctly...)

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js
(right):

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:77: });
On 2016/05/11 01:54:05, yhirano wrote:
> +indent

Done.

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:78: },
'fetch blob: URL');
On 2016/05/11 01:54:04, yhirano wrote:
> +indent

Done.

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:84:
t.unreached_func('fetching non-existent blob: URL must fail'),
On 2016/05/11 01:54:05, yhirano wrote:
> -indent

Done.

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:85:
function() {});
On 2016/05/11 01:54:05, yhirano wrote:
> -indent

Done.

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:86: },
'fetch non-existent blob: URL');
On 2016/05/11 01:54:05, yhirano wrote:
> +indent

Done.

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
File
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/thorough/scheme-blob.js
(right):

https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/thorough/scheme-blob.js:6:
// Only [Exposed=(Window,DedicatedWorker,SharedWorker)].
On 2016/05/11 01:54:05, yhirano wrote:
> Could you define an empty TEST_TARGETS here and move
> |executeTests(TEST_TARGETS)| back to the last?

Done.

[Rick Byers, 2016-05-12 01:07:08]

On 2016/05/11 21:52:42, jsbell wrote:
> rbyers@ - can you API_OWNERS review?

third_party/WebKit/LayoutTests/virtual/stable LGTM

[yhirano, 2016-05-12 04:48:24]

On 2016/05/11 21:52:42, jsbell wrote:
> rbyers@ - can you API_OWNERS review?
> 
> yhirano@ - thanks for the detailed formatting feedback! Did we ever figure out
> clang-format settings for that style? (since I suspect I'll never get emacs to
> do it correctly...)
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> File third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js
> (right):
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:77: });
> On 2016/05/11 01:54:05, yhirano wrote:
> > +indent
> 
> Done.
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:78: },
> 'fetch blob: URL');
> On 2016/05/11 01:54:04, yhirano wrote:
> > +indent
> 
> Done.
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:84:
> t.unreached_func('fetching non-existent blob: URL must fail'),
> On 2016/05/11 01:54:05, yhirano wrote:
> > -indent
> 
> Done.
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:85:
> function() {});
> On 2016/05/11 01:54:05, yhirano wrote:
> > -indent
> 
> Done.
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/fetch.js:86: },
> 'fetch non-existent blob: URL');
> On 2016/05/11 01:54:05, yhirano wrote:
> > +indent
> 
> Done.
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
> File
>
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/thorough/scheme-blob.js
> (right):
> 
>
https://codereview.chromium.org/1908263002/diff/80001/third_party/WebKit/Layo...
>
third_party/WebKit/LayoutTests/http/tests/fetch/script-tests/thorough/scheme-blob.js:6:
> // Only [Exposed=(Window,DedicatedWorker,SharedWorker)].
> On 2016/05/11 01:54:05, yhirano wrote:
> > Could you define an empty TEST_TARGETS here and move
> > |executeTests(TEST_TARGETS)| back to the last?
> 
> Done.

[yhirano, 2016-05-12 04:49:14]

yhirano@chromium.org changed reviewers:
+ falken@chromium.org

[yhirano, 2016-05-12 04:49:15]

> Did we ever figure out clang-format settings for that style? (since I suspect
I'll never get emacs to do it correctly...)

falken@, do we have any tool support for the service worker laytout tests style?

[falken, 2016-05-12 05:02:14]

On 2016/05/12 04:49:15, yhirano wrote:
> > Did we ever figure out clang-format settings for that style? (since I
suspect
> I'll never get emacs to do it correctly...)
> 
> falken@, do we have any tool support for the service worker laytout tests
style?

Unfortunately, no.

[jsbell, 2016-05-12 16:34:52]

The CQ bit was checked by jsbell@chromium.org

[jsbell, 2016-05-12 16:34:53]

The patchset sent to the CQ was uploaded after l-g-t-m from
haraken@chromium.org, tommi@chromium.org, wolenetz@chromium.org,
yhirano@chromium.org
Link to the patchset: https://codereview.chromium.org/1908263002/#ps100001
(title: "test cleanup")

[commit-bot: I haz the power, 2016-05-12 16:35:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1908263002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1908263002/100001

[commit-bot: I haz the power, 2016-05-12 16:39:07]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard. Regardless, the overloads should
never execute in Service Workers as those types (MediaStream,
MediaSource) are not exposed in that context.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard. Regardless, the overloads should
never execute in Service Workers as those types (MediaStream,
MediaSource) are not exposed in that context.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460
==========

[commit-bot: I haz the power, 2016-05-12 16:39:08]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-05-12 16:40:37]

Description was changed from

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no 
longer present in the standard. Regardless, the overloads should
never execute in Service Workers as those types (MediaStream,
MediaSource) are not exposed in that context.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460
==========

to

==========
Don't expose URL.createObjectURL and revokeObjectURL to Service Workers

Per FileAPI[1] which was updated based on Service Worker discussion[2],
minting blob URLs should not be possible within SWs.

A bug has been filed against MediaSource[3] to update the IDL there
(following the same logic), and mediastream's createObjectURL is no
longer present in the standard. Regardless, the overloads should
never execute in Service Workers as those types (MediaStream,
MediaSource) are not exposed in that context.

Note that the [Exposed] attributes are applied directly to the
interface members as they appear to be ignored on the partial
interface. (Bug?)

[1] https://w3c.github.io/FileAPI/#creating-revoking
[2] https://github.com/slightlyoff/ServiceWorker/issues/688
[3] https://github.com/w3c/media-source/issues/67

Intent Thread:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/HuA7Ng9U0oc/CYvfMoey...

BUG=604951,608460

Committed: https://crrev.com/9e232c512c962a7aecc85e07f7a48a4735dc7a25
Cr-Commit-Position: refs/heads/master@{#393271}
==========

[commit-bot: I haz the power, 2016-05-12 16:40:39]

Patchset 6 (id:??) landed as
https://crrev.com/9e232c512c962a7aecc85e07f7a48a4735dc7a25
Cr-Commit-Position: refs/heads/master@{#393271}