W3C home > Mailing lists > Public > public-html-testsuite@w3.org > November 2010

Assumptions and Test metadata (was: Re: Canvas Test Submission approval/feedback request)

From: Geoffrey Sneddon <gsneddon@opera.com>
Date: Sat, 06 Nov 2010 13:19:03 +0000
Message-ID: <4CD555C7.8010005@opera.com>
To: James Graham <jgraham@opera.com>
CC: Kris Krueger <krisk@microsoft.com>, Aryeh Gregor <Simetrical+w3c@gmail.com>, Philip Taylor <pjt47@cam.ac.uk>, Anne van Kesteren <annevk@opera.com>, "'public-html-testsuite@w3.org' (public-html-testsuite@w3.org)" <public-html-testsuite@w3.org>
On 06/10/10 09:14, James Graham wrote:
> On Wed, 6 Oct 2010, Kris Krueger wrote:
>> Really glad to see all the participation and communication.
>> As the appointed testing task force lead, james is correct according
>> to our current charter from the co-chairs.
>> That said this is a simple valuable use case for svg and canvas....
>> This is also the reason Microsoft submitted this case to make it clear
>> on how canvas should be built in an interoperable way.
>> Note that safari, chrome and IE9 supports this today, so it's indeed
>> possible to be implemented by more than one vendor.
> After further discussions, I'm not sure my original position is tenable
> without causing undue inconvenience in writing testcases. Nevertheless
> where optional features exist we need some critera to determine whether
> we should include them in the official testsuite. A reasonable policy
> might be "features that all four major engines are committed to
> implementing are OK to test without a precondition check". It seems that
> SVG-in-img (and hence SVG-in-drawImage) meet this criterion (whereas,
> say, PDF-in-<img> would not). Therefore I no longer have a principled
> objection to the substance of this testcase.

I still think the proposed solution is a viable one (with the one 
correction as pointed out by Simon), but we probably ought to have more 
agreement about what we should assume about UAs. At the very least, it 
seems reasonable to assume support for GIF, PNG, and SVGT.

That said, I think we do need more metadata than we currently have; I'd 
be perfectly happy to go with a solution similar to CSS 2.1's flags, see 
<http://wiki.csswg.org/test/css2.1/format#requirement-flags>. This would 
mean that any other format would be flagged with "may".

It also seems worthwhile to bring up video here: with no agreement on a 
common format that we can assume support for, we seem to have two basic 
choices: label all video tests as "may" (for we can't rely upon the 
format being supported) or add two new flags, "h264" and "webm" (do we 
still need to test Ogg/Vorbis/Theora?) which require support for those 
formats. I guess the latter option is the only realistic one.

Geoffrey Sneddon  Opera Software
Received on Saturday, 6 November 2010 13:19:40 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 15:49:37 UTC