RE: RfR: WebSocket Api tests; deadline October 28th

Thanks Art, 

 I updated the tests, updated status.html and moved all of Ms2Ger and Microsoft tests to the approved folder.

Opera will be submitting some additional WebSocket tests.  Once Opera's tests have been reviewed and approved.  We should ask for implementation report to ensure we have at least two passing browsers.  Once this is done we should be all set to move the spec from CR to PR.

-Kris

-----Original Message-----
From: Arthur Barstow [mailto:art.barstow@nokia.com] 
Sent: Thursday, October 25, 2012 7:40 PM
To: Kris Krueger
Cc: public-webapps-testsuite@w3.org
Subject: Re: RfR: WebSocket Api tests; deadline October 28th

On 10/20/12 2:03 PM, ext Kris Krueger wrote:
>
> This is a Request for Review [RfR] for all of the Web Socket tests 
> that have been submitted by Ms2ger and Microsoft for the WebSocket API
>  spec:
>
> The tests have been updated to accommodate the specification changes 
> dealing with ArrayBufferView and unpaired surrogates.
>
> http://www.w3c-test.org/webapps/WebSockets/tests/submissions/Microsoft
> /
>
> http://www.w3c-test.org/webapps/WebSockets/tests/submissions/Ms2ger/
>

Hi All,

I reviewed Ms2ger's tests and they look OK to me.

I reviewed 46 of Microsoft's tests (Create-*, Close-* and
Secure-Close-*) and I don't expect to have time to review the other tests before the deadline. The tests I reviewed look OK to me,  although I have a fewcomments below.

Perhaps someone that doesn't have time to review them all can look at the others i.e. Send-* and Secure-Send-*?Would someone please commit to review these?

-AB

* Close-1000 and Close-1000-reason - the description of the first tests
("testOpen") don't appear to be accurate, forexample those tests don't actually check readyState at all.

* Create-Secure-valid-url-protocol-string - this gist of this test is to see if readyState is set to 1 in the open callback but the description indicates otherwise. Coverage wise, this test seems to effectively be identical to Create-Secure-valid-url-protocol-setCorrectly.

* Secure-Close-1000-verify-code - seems like the testClose descriptionshould reflect the two assert_equals i.e. evt.code == 1000 and evt.reason== "Clean Close" rather than mention readyState and wasClean states.

* Secure-Close-1005* - it would probably be helpful to add a reference to RFC6455 to these two tests. (BTW, Chrome 22.0.1229.94 returns 1006for the verify-code test and not 1005 as the test expects).

> If you have any comments, please send them by October 28th.
>
> If you review any set of the tests and find no issues, please state 
> that as a reply to this RfR (so we can get a sense of whether or not 
> anyone reviewed the tests).
>
> In the absence of any comments, these tests will be considered Approved.
>
> -Kris
>
> [RfR] http://www.w3.org/2008/webapps/wiki/Approval
> <http://www.w3.org/2008/webapps/wiki/Approval>
>

Received on Tuesday, 30 October 2012 11:06:17 UTC