W3C home > Mailing lists > Public > public-html-testsuite@w3.org > December 2012

RE: Migrate your submissions

From: Kris Krueger <krisk@microsoft.com>
Date: Sat, 22 Dec 2012 01:21:08 +0000
To: James Graham <jgraham@opera.com>, David Carlisle <davidc@nag.co.uk>
CC: "public-html-testsuite@w3.org" <public-html-testsuite@w3.org>
Message-ID: <0A605FCDA3A4DC45A98B0DDD1C5351A70390E24C@TK5EX14MBXW601.wingroup.windeploy.ntdev.microsoft.com>
Though do note that the HTML5 spec has no normative references to how mathml actually renders.
The spec only covers the parsing of MathML.


-----Original Message-----
From: James Graham [mailto:jgraham@opera.com] 
Sent: Thursday, December 20, 2012 11:58 AM
To: David Carlisle
Cc: public-html-testsuite@w3.org
Subject: Re: Migrate your submissions

On Thu, 20 Dec 2012, David Carlisle wrote:

> On 20/12/2012 17:35, Robin Berjon wrote:
>> Hi all,
>> if you've made a test submission that has not been processed in the
>> past, please read this.
>> I've done a lot of migrating submissions into pull requests today.
>> This includes the following:
>> DavidCarlisle
> Thanks.
> I originally made a couple of "test" submissions for mathml
> support, for parsing and for a minimal rendering tests (just checking
> that the rendering was different from <span> more or less.
> The parsing tests were approved, the rendering tests were (not
> unreasonably) held up as it wasn't clear what the process should be for
> rendering tests (despite the fact that the foreign-content tests that
> were originally approved had several visual svg rendering tests).
> It would be good to get more tests in but is the policy any clearer now?
> Are tests of rendering allowed? If so is there a specific part of the
> test suite for them?

Tests for rendering are allowed. Assuming the use of javascript for the 
tests is impossible or very undesirable, reftests are strongly preferred 
(rendering tests MUST be reftests where possible, if you will). I don't 
recall if this was a problem with your submission, but tests, especially 
those that cannot be tested automatically, must have a clear pass 
condition so that they can be run by someone with no knowledge of the 
technology under test in a few seconds per test.
Received on Saturday, 22 December 2012 01:21:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:14:31 UTC