W3C home > Mailing lists > Public > public-web-perf@w3.org > January 2012

RE: IETC tests for PageVisibility

From: Karen Anderson (IE) <Karen.Anderson@microsoft.com>
Date: Wed, 25 Jan 2012 21:38:31 +0000
To: Tony Gentilcore <tonyg@google.com>
CC: Jatinder Mann <jmann@microsoft.com>, "public-web-perf@w3.org" <public-web-perf@w3.org>
Message-ID: <18D01621E7D58B4CA8BA9DDBD0D8B12C0388441F@TK5EX14MBXC286.redmond.corp.microsoft.com>
Yes, I now remember the conversation in the context of additional prefixed attributes.  Sorry.

The TC is interesting b/c even MS is prefixed.  So it is not fair to check against the prefix for MS only.  So it does have logic to iterate over the known prefixes as I suggested below.

I plan to have a drop of tests available either EOW or early next week...hopefully in time for all to review prior to our call.

Thanks,
Karen

-----Original Message-----
From: Tony Gentilcore [mailto:tonyg@google.com] 
Sent: Wednesday, January 25, 2012 1:20 AM
To: Karen Anderson (IE)
Cc: Jatinder Mann; public-web-perf@w3.org
Subject: Re: IETC tests for PageVisibility

On Tue, Jan 24, 2012 at 11:20 PM, Karen Anderson (IE) <Karen.Anderson@microsoft.com> wrote:
> Yes.  Unfortunately, the Testing Center is only updated at the same schedule as public available builds.  So, the next opportunity will be Beta (soon).  I will make sure it's on that wave.
>
> On a side note...when we published Navigation Timing tests, from what I can tell they have always been non-vendor prefixed.  It seems that all implementations of Page Visibility currently are prefixed...how do we want to handle this?  The main test harness does not do a lookup, but I think that would be a valuable addition.  For example, in NT's wp_test function where the namespace is checked, we could have a similar function that checks document.visibilityState, and if undefined, then iterates over webkit, moz, o, and ms.

Up to you about the way to handle this on IETC, but when we discussed this previously we came to the conclusion that W3C tests should not pass for vendor prefixed implementations. A vendor-prefixed implementation, even if otherwise compliant, is not fully interoperable. That being said, it is useful to structure the tests in such a way that it is simple to modify one line to get them to recognize the prefixed implementation. That allows us browser developers to use them while the interface is still under development.

For bonus points, we could structure the tests to recognize the vendor-prefixed implementations, perform all of the tests (instead of failing fast), and output FAIL for each with a note in the error message that it would pass except for the prefix.

>
> For now I will keep this self contained, but I think we might hit this again in other specs, so at some point this should really live in a more centralized location.
>
> Thoughts?
> Karen
>
>
> -----Original Message-----
> From: Tony Gentilcore [mailto:tonyg@google.com]
> Sent: Tuesday, January 24, 2012 2:46 PM
> To: Jatinder Mann
> Cc: public-web-perf@w3.org
> Subject: Re: IETC tests for PageVisibility
>
> Excellent. Thanks :)
>
> On Tue, Jan 24, 2012 at 7:59 PM, Jatinder Mann <jmann@microsoft.com> wrote:
>> Tony,
>>
>> We wrote these tests with the intention of making them available as a part of the Page Visibility W3C test suite. I believe Karen is working on updating them to work in the w3c test harness. As the Page Visibility spec has been relatively stable, I am hoping to submit these tests soon so we can move the spec status forward.
>>
>> Yes, the constant tests need to be deleted as that portion of the spec has been updated.
>>
>> Thanks,
>> Jatinder
>>
>> -----Original Message-----
>> From: Tony Gentilcore [mailto:tonyg@google.com]
>> Sent: Tuesday, January 24, 2012 11:20 AM
>> To: public-web-perf@w3.org
>> Subject: IETC tests for PageVisibility
>>
>> It came to my attention there are some PageVisibility tests [1] [2] [3] in the IE Test Center suite which look like they'd be good candidates to contribute to http://w3c-test.org/webperf/tests/.
>> Currently WebKit fails because they are incorrectly testing for the PAGE_HIDDEN/PAGE_VISIBLE string constants which have been removed from the latest version of the spec [4].
>>
>> Just curious what the plans are for these tests.
>>
>> -Tony
>>
>> [1]
>> http://samples.msdn.microsoft.com/ietestcenter/WebPerf/PageVisibility
>> /
>> test_api_exists_attributes.htm [2]
>> http://samples.msdn.microsoft.com/ietestcenter/WebPerf/PageVisibility
>> / test_read_only_conformance_attributes.htm
>> [3]
>> http://samples.msdn.microsoft.com/ietestcenter/WebPerf/PageVisibility
>> /
>> test_iframe_attributes.htm [4]
>> https://dvcs.w3.org/hg/webperf/raw-file/tip/specs/PageVisibility/Over
>> v iew.html#sec-document-visibility-interface
>>
>>
>>
>
>
>
Received on Wednesday, 25 January 2012 21:39:14 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:04:32 UTC