Refactor WebUSB LayoutTests to separate out the Mojo service mocks

Refactor WebUSB LayoutTests to separate out the Mojo service mocks

In preparation for being able to upstream the existing WebUSB tests to
the Web Platform Tests repository this patch splits out the Mojo
service mocks into webusb-test.js. This file is a polyfil for a
fictional "WebUSB Test API" that could in theory be implemented by any
browser (an idea borrowed from a proposed plan for Web Bluetooth).

BUG=705734
Review-Url: https://codereview.chromium.org/2775323003
Cr-Commit-Position: refs/heads/master@{#460978}
Committed: https://chromium.googlesource.com/chromium/src/+/e1edfbbb4b3a192e210c031628e4f12aa93057cd

[Reilly Grant (use Gerrit), 2017-03-27 22:51:41]

The CQ bit was checked by reillyg@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-03-27 22:52:07]

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

[Reilly Grant (use Gerrit), 2017-03-27 22:52:48]

reillyg@chromium.org changed reviewers:
+ ortuno@chromium.org

[Reilly Grant (use Gerrit), 2017-03-27 22:52:49]

Please take a look. Since I anticipated this would be easier for WebUSB than Web
Bluetooth because of the smaller API surface I gave your Test API polyfil idea a
try and this is the resulting patch.

[commit-bot: I haz the power, 2017-03-27 23:51:47]

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

[commit-bot: I haz the power, 2017-03-27 23:51:48]

Dry run: This issue passed the CQ dry run.

[ortuno, 2017-03-29 00:12:30]

Thanks for doing this! This looks exactly like what I had in mind :)

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/usb/resources/webusb-test.js (right):

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:16: //   DOMString
addFakeDevice(FakeUSBDeviceInit deviceInit);
We made our whole API promised based so that we could implement it on top of
real devices. I'm not sure how USB works but do you think we would still be able
to synchronously add fake devices if we were implementing this on top of real
drivers. If you aren't planning on implementing this outside of content shell
then I guess it doesn't really matter.

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:459:
gin.define('WebUSB Test Mocks', [
This seems similar to what mojo_helpers.js has. What do you think of moving it
to testharnessreport.js For context: testharnessreport.js is a file that each UA
has to provide which includes UA-specific methods.

[Reilly Grant (use Gerrit), 2017-03-29 00:29:54]

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/usb/resources/webusb-test.js (right):

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:16: //   DOMString
addFakeDevice(FakeUSBDeviceInit deviceInit);
On 2017/03/29 00:12:30, ortuno wrote:
> We made our whole API promised based so that we could implement it on top of
> real devices. I'm not sure how USB works but do you think we would still be
able
> to synchronously add fake devices if we were implementing this on top of real
> drivers. If you aren't planning on implementing this outside of content shell
> then I guess it doesn't really matter.

I looked over the Web Bluetooth Test API and the guarantees you make are
stronger because the Promise is only resolved once the renderer has been
informed of the new device, while this method simply kicks off the creation of a
fake device and so could remain synchronous even if it was calling into a the
browser process via a Mojo interface.

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:459:
gin.define('WebUSB Test Mocks', [
On 2017/03/29 00:12:30, ortuno wrote:
> This seems similar to what mojo_helpers.js has. What do you think of moving it
> to testharnessreport.js For context: testharnessreport.js is a file that each
UA
> has to provide which includes UA-specific methods.

I duplicated the logic in mojo-helpers.js to avoid pulling in more of the Mojo
bindings framework than I needed to. I could move some of this into
testharnessreport.js but I like how putting it here makes the whole polyfil
self-contained. I plan to move this file to platform-fakes/webusb-test.js as
part of the upstreaming.

[ortuno, 2017-03-29 20:56:00]

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/usb/resources/webusb-test.js (right):

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:459:
gin.define('WebUSB Test Mocks', [
On 2017/03/29 at 00:29:53, Reilly Grant wrote:
> On 2017/03/29 00:12:30, ortuno wrote:
> > This seems similar to what mojo_helpers.js has. What do you think of moving
it
> > to testharnessreport.js For context: testharnessreport.js is a file that
each UA
> > has to provide which includes UA-specific methods.
> 
> I duplicated the logic in mojo-helpers.js to avoid pulling in more of the Mojo
bindings framework than I needed to. I could move some of this into
testharnessreport.js but I like how putting it here makes the whole polyfil
self-contained. I plan to move this file to platform-fakes/webusb-test.js as
part of the upstreaming.

Makes sense. Another question, mostly to figure out what obstacles do we have
for running this in regular chrome, can we use gin outside of content shell?

[Reilly Grant (use Gerrit), 2017-03-29 21:04:14]

reillyg@chromium.org changed reviewers:
+ yzshen@chromium.org

[Reilly Grant (use Gerrit), 2017-03-29 21:04:14]

+yzshen@ for the Mojo interface override in normal Chrome question.

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/usb/resources/webusb-test.js (right):

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:459:
gin.define('WebUSB Test Mocks', [
On 2017/03/29 20:56:00, ortuno wrote:
> On 2017/03/29 at 00:29:53, Reilly Grant wrote:
> > On 2017/03/29 00:12:30, ortuno wrote:
> > > This seems similar to what mojo_helpers.js has. What do you think of
moving
> it
> > > to testharnessreport.js For context: testharnessreport.js is a file that
> each UA
> > > has to provide which includes UA-specific methods.
> > 
> > I duplicated the logic in mojo-helpers.js to avoid pulling in more of the
Mojo
> bindings framework than I needed to. I could move some of this into
> testharnessreport.js but I like how putting it here makes the whole polyfil
> self-contained. I plan to move this file to platform-fakes/webusb-test.js as
> part of the upstreaming.
> 
> Makes sense. Another question, mostly to figure out what obstacles do we have
> for running this in regular chrome, can we use gin outside of content shell?

Last time I checked we were not able to use this interface outside of
content_shell when run with --run-layout-tests but I think it makes sense to add
a flag that enables Mojo interface overrides in any version of Chrome.

[ortuno, 2017-03-29 21:12:57]

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
File third_party/WebKit/LayoutTests/usb/resources/webusb-test.js (right):

https://codereview.chromium.org/2775323003/diff/1/third_party/WebKit/LayoutTe...
third_party/WebKit/LayoutTests/usb/resources/webusb-test.js:459:
gin.define('WebUSB Test Mocks', [
On 2017/03/29 at 21:04:14, Reilly Grant wrote:
> On 2017/03/29 20:56:00, ortuno wrote:
> > On 2017/03/29 at 00:29:53, Reilly Grant wrote:
> > > On 2017/03/29 00:12:30, ortuno wrote:
> > > > This seems similar to what mojo_helpers.js has. What do you think of
moving
> > it
> > > > to testharnessreport.js For context: testharnessreport.js is a file that
> > each UA
> > > > has to provide which includes UA-specific methods.
> > > 
> > > I duplicated the logic in mojo-helpers.js to avoid pulling in more of the
Mojo
> > bindings framework than I needed to. I could move some of this into
> > testharnessreport.js but I like how putting it here makes the whole polyfil
> > self-contained. I plan to move this file to platform-fakes/webusb-test.js as
> > part of the upstreaming.
> > 
> > Makes sense. Another question, mostly to figure out what obstacles do we
have
> > for running this in regular chrome, can we use gin outside of content shell?
> 
> Last time I checked we were not able to use this interface outside of
content_shell when run with --run-layout-tests but I think it makes sense to add
a flag that enables Mojo interface overrides in any version of Chrome.

I talked to rockot and he mentioned that we it'll be straightforward to do this.
I was just curious if gin was something we could already use outside of Layout
Tests.

[ortuno, 2017-03-30 22:19:07]

lgtm!

[Reilly Grant (use Gerrit), 2017-03-30 23:26:24]

The CQ bit was checked by reillyg@chromium.org

[commit-bot: I haz the power, 2017-03-30 23:26:53]

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, 2017-03-31 01:22:11]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1490916384542200, "parent_rev":
"ec4a8e3a9d0000eaaefe7a02bd2a405bf510d49d", "commit_rev":
"e1edfbbb4b3a192e210c031628e4f12aa93057cd"}

[commit-bot: I haz the power, 2017-03-31 01:23:00]

Description was changed from

==========
Refactor WebUSB LayoutTests to separate out the Mojo service mocks

In preparation for being able to upstream the existing WebUSB tests to
the Web Platform Tests repository this patch splits out the Mojo
service mocks into webusb-test.js. This file is a polyfil for a
fictional "WebUSB Test API" that could in theory be implemented by any
browser (an idea borrowed from a proposed plan for Web Bluetooth).

BUG=705734
==========

to

==========
Refactor WebUSB LayoutTests to separate out the Mojo service mocks

In preparation for being able to upstream the existing WebUSB tests to
the Web Platform Tests repository this patch splits out the Mojo
service mocks into webusb-test.js. This file is a polyfil for a
fictional "WebUSB Test API" that could in theory be implemented by any
browser (an idea borrowed from a proposed plan for Web Bluetooth).

BUG=705734

Review-Url: https://codereview.chromium.org/2775323003
Cr-Commit-Position: refs/heads/master@{#460978}
Committed:
https://chromium.googlesource.com/chromium/src/+/e1edfbbb4b3a192e210c031628e4...
==========

[commit-bot: I haz the power, 2017-03-31 01:23:04]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/chromium/src/+/e1edfbbb4b3a192e210c031628e4...

[ortuno, 2017-03-31 05:02:06]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/2791573002/ by ortuno@chromium.org.

The reason for reverting is: Breaks leak bot:

https://uberchromegw.corp.google.com/i/chromium.webkit/builders/WebKit%20Linu...

* usb/test-polyfil.html
* usb/usb-connection-event.html
* usb/usb.html
* usb/usbDevice-iframe.html
* usb/usbDevice.html

seem to be leaking :(.