Re: Urgent need for EME test case review

On Thu, Jul 28, 2016 at 2:49 PM, Matt Wolenetz <wolenetz@google.com> wrote:

> In case you all are not aware, @ddorwin has low availability this week. I
> understand that he's expecting to be back in the office on Monday. I
> believe he mentioned this during previous editors' sync, but I don't see
> such in the minutes text.
>

​I feel that we should revisit our 2 August deadline ...​



>
> On Thu, Jul 28, 2016 at 10:38 AM, Mark Watson <watsonm@netflix.com> wrote:
>
>> All,
>>
>> I have broken down the test cases I have created into 7 Pull Requests:
>>
>> https://github.com/w3c/web-platform-tests/pull/3355
>> https://github.com/w3c/web-platform-tests/pull/3359
>> https://github.com/w3c/web-platform-tests/pull/3360
>> https://github.com/w3c/web-platform-tests/pull/3361 (this has a lint
>> problem which I will fix)
>> https://github.com/w3c/web-platform-tests/pull/3362
>> https://github.com/w3c/web-platform-tests/pull/3363
>> https://github.com/w3c/web-platform-tests/pull/3364
>>
>> You can see them all here
>> <https://github.com/w3c/web-platform-tests/labels/encrypted-media>.
>>
>> I will close the huge one without merging.
>>
>> Remaining items on the list for Sukhmal and I are:
>>
>>    - Migrate further Google tests (at least those which do no rely in
>>    Blink-specific window.internals)
>>    - Create WebM versions (though I would prefer someone else to pick
>>    this up)
>>    - Persistent license tests
>>    - Update the content for multi-key / multi-session test to remove
>>    discontinuity
>>    - Sequential multi-key / single session test
>>    - Review Firefox results and check if any failures are due to test
>>    bugs / polyfillable browser bugs
>>    - Run on Edge and review results to check if any failures are due to
>>    test bugs / polyfillable browser bugs
>>    - Run on CastTV and review results to check if any failures are due
>>    to test bugs / polyfillable browser bugs
>>
>> We will not be able to run on Safari / IE unless we polyfill the entire
>> API on top of their prefixed versions. This is possible, and not all that
>> much work, but I am not sure how informative it would be.
>>
>> FWIW, the test results for Chrome (Canary) (runner-results.json) and
>> Firefox 47 (runner-results(1).json) (both on Mac) are attached. As noted
>> above, I have not reviewed the Firefox ones yet.
>>
>> Btw, I see the test results JSON lists the test file and the test name.
>> Presently, the only difference between the "real" tests and the "polyfill"
>> ones is the test file path - the test names are the same. If we need to
>> also reflect that in the test name, I can do that. LMK.
>>
>> ...Mark
>>
>>
>>
>>
>> On Wed, Jul 27, 2016 at 1:28 PM, Mark Watson <watsonm@netflix.com> wrote:
>>
>>> All,
>>>
>>> The previous pull request has become very large. I will break it up into
>>> a sequence of smaller ones. To facilitate this, I need to merge one PR that
>>> establishes the latest framework. The other PRs can then be reviewed in
>>> parallel.
>>>
>>> So, the first of this sequence is
>>> https://github.com/w3c/web-platform-tests/pull/3355/files
>>>
>>> I URGENTLY need review of this PR. If we are to meet our deadline of
>>> August 2, I presume we need the tests running this week and I will have
>>> limited time to work on this tomorrow and Friday. I don't think there is
>>> much point in attempting to meet the August 2 deadline if noone is able to
>>> review the test PRs.
>>>
>>> ...Mark
>>>
>>>
>>>
>>> On Mon, Jul 25, 2016 at 5:29 PM, Mark Watson <watsonm@netflix.com>
>>> wrote:
>>>
>>>> All,
>>>>
>>>> Update for today: My Pull Request (
>>>> https://github.com/w3c/web-platform-tests/pull/3324) now has updated
>>>> mp4 content and consolidated content metadata. Some re-arrangement is still
>>>> needed to support the AV multi-key case (separate audio and video keys). I
>>>> expect to get content tomorrow with multiple keys within one video stream.
>>>> I'll continue to add tests tomorrow.
>>>>
>>>> The space of combinations of content type, key system and init data
>>>> type is quite large and not fully covered (or all that well organized in
>>>> the code). Some tests just need to discover a supported combination, others
>>>> are constrained by what the DRM server supports as well as what the client
>>>> component supports. I'll also look at how we can organize this better.
>>>>
>>>> As before, review of the PR is welcome!
>>>>
>>>> ...Mark
>>>>
>>>>
>>>>
>>>> On Mon, Jul 25, 2016 at 11:21 AM, Paul Cotton <
>>>> Paul.Cotton@microsoft.com> wrote:
>>>>
>>>>> > I will be working on the other tests on and off early next week. If
>>>>> anyone else is able to help, please coordinate with me.
>>>>>
>>>>>
>>>>>
>>>>> I am moving this discussion [1] to public-html-media@w3.org so that
>>>>> we touch as many HME WG members as possible.
>>>>>
>>>>>
>>>>>
>>>>> Mark and others:  Please use this thread for any future updates on the
>>>>> status of EME testing.
>>>>>
>>>>>
>>>>>
>>>>> /paulc
>>>>>
>>>>>
>>>>>
>>>>> [1]
>>>>> https://lists.w3.org/Archives/Public/public-hme-editors/2016Jul/0056.html
>>>>>
>>>>>
>>>>>
>>>>> *From:* Mark Watson [mailto:watsonm@netflix.com]
>>>>> *Sent:* Friday, July 22, 2016 5:25 PM
>>>>> *To:* Paul Cotton <Paul.Cotton@microsoft.com>
>>>>> *Cc:* Jerry Smith (WPT) <jdsmith@microsoft.com>; Greg Rutz <
>>>>> G.Rutz@cablelabs.com>; David Dorwin <ddorwin@google.com>; Matthew
>>>>> Wolenetz <wolenetz@google.com> (wolenetz@google.com) <
>>>>> wolenetz@google.com>; Philippe Le Hegaret (plh@w3.org) <plh@w3.org>;
>>>>> Francois Daoust <fd@w3.org>; public-hme-editors@w3.org; Iraj Sodagar <
>>>>> irajs@microsoft.com>; John Simmons <johnsim@microsoft.com>; Sukhmal
>>>>> Kommidi <skommidi@netflix.com>
>>>>> *Subject:* Re: DRM Today-based test case for EME
>>>>>
>>>>>
>>>>>
>>>>> All,
>>>>>
>>>>>
>>>>>
>>>>> I migrated a handful of the Google tests:
>>>>> https://github.com/w3c/web-platform-tests/pull/3324
>>>>>
>>>>>
>>>>>
>>>>> Currently, the "drm-events" test fails because the DRM server is
>>>>> unaware of the key id that is used (and I am unaware of the key id that it
>>>>> does know). I'm working with Greg on this.
>>>>>
>>>>>
>>>>>
>>>>> I noticed the following whilst doing this work:
>>>>>
>>>>>    - We probably need to organize the content metadata (key ids,
>>>>>    keys, MIME types etc.) into one place - presently they are scattered around
>>>>>    in a confusing way
>>>>>    - The idea of auto-generating the HTML stubs needs some more
>>>>>    thought - they are presently a bit different for each test
>>>>>    - Some of the Google tests look (to me at least) like they might
>>>>>    be Chrome-specific. Specifically, the lifetime tests use window.internals
>>>>>    and I am not sure if this is cross-platform. Likewise, I do not know if the
>>>>>    garbage collection stuff is cross-platform or not.
>>>>>    - We need some additional mp4 content files (I am in touch with
>>>>>    Greg about this):
>>>>>
>>>>>
>>>>>    - Encrypted audio
>>>>>       - Video with multiple keys
>>>>>
>>>>> I will be working on the other tests on and off early next week. If
>>>>> anyone else is able to help, please coordinate with me.
>>>>>
>>>>>
>>>>>
>>>>> Sukhmal is working on persistent-usage-record and persistent-license
>>>>> tests, which should be ready next week.
>>>>>
>>>>>
>>>>>
>>>>> ...Mark
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>

Received on Thursday, 28 July 2016 23:51:10 UTC