Re: First chrome wpt test added

On Mon, Oct 30, 2017 at 1:33 AM, Paul Adenot <padenot@mozilla.com> wrote:

> Thanks Raymond.
>
> It looks like Firefox does not pass some assertions, I'll try to have a
> look.
>

I think they're relatively minor.  The min/max values are set to +/-
infinity on Firefox, but most positive (negative) single-float on Chrome.
I think the spec did say the nominal range was +/- infinity, exclusive, and
the latest version does use the most positive (negative) single float
values.

>
> Tests from Mozilla are still in our repo [0], and will be merged in the
> next few days (not sure when). They cover AnalyserNode, mostly.
> Additionally, your tests should be merged into our repo, and should start
> running (and I should fix Firefox so they fully pass, but I don't see
> myself having enough time right now).
>

You want to merge these tests back to the mozilla repo? I'm fine with that,
but we're planning on removing these constant source node tests from
chrome's tests, if possible.


>
> Additionally, I'm mentoring students from a french university to convert
> more Mozilla tests, and I'm planning to convert some more myself.
>

Awesome!

One final detail.  As I mentioned, we have a wrapper  (Audit) around
testharness and I would like to know the way forward?  I would certainly
prefer all tests to use the wrapper (because the failure messages are quite
detailed), but we can certainly live with a mixture of testharness test and
Chrome Audit.  I don't won't to convert our tests to plain testharness test.



> Paul.
>
> [0]: https://hg.mozilla.org/mozilla-central/file/tip/
> testing/web-platform/tests/webaudio/the-audio-api/the-
> analysernode-interface
>
> On Fri, Oct 27, 2017 at 9:28 PM, Raymond Toy <rtoy@google.com> wrote:
>
>> We've landed our first chrome wpt test which is for ConstantSourceNode,
>> because it's relatively self-contained.
>>
>> You can see the results at https://wpt.fyi/webaudio/ch
>> rome/the-constantsourcenode-interface
>>
>> As discussed briefly on the teleconf, we do want to move these to the
>> right place eventually, but to do that we want to coordinate with all the
>> browser vendors first.
>>
>> Thus, we'd like the vendors to take a look at the tests and see what you
>> think and whether we should move forward with these.  In particular, Chrome
>> has a wrapper around the wpt test methods to make it easier to write and
>> more informative when the tests fail.
>>
>> The relevant change is https://github.com/w3c/web-
>> platform-tests/commit/fabf91dc28ca3e2ae83099b6b169fc16ce657b06
>>
>
>