Re: CSS Grid Layout tests suite

Le 2014-10-27 06:14, Manuel Rego Casasnovas a écrit :
> Hi,
> 
> thanks for the reply.
> 
> On 25/10/14 20:34, Gérard Talbot wrote:
>>> [1] http://test.csswg.org/shepherd/search/spec/css-grid-1/
>> 
>> Rego,
>> 
>> I can not view the existing Grid Layout 1 tests right now with 
>> Shepherd.
> 
> I don't know why. The link above works fine for me, even if I'm not
> logged in.

Manuel,

There was a problem (involving viewing tests and source code within 
Shepherd application) that Peter Linss fixed yesterday.

> 
>> I can make some comments on 1 test (randomly chosen). I understand you
>> are not the author of such test... and it appears that those tests 
>> were
>> based on an outdated version of the spec.
> 
> Yes, you're right, I'm not the author and they're outdated.
> Thanks for the review, I agree with all of your comments.
> 
> However, as I'm not the author, I don't know what's the best approach 
> to
> carry on.

Normally, someone (the test reviewer) would try to contact/email the 
original author so that these issues would be fixed, corrected.

> I'm just wondering if it'd be a better idea to start from scratch with
> new tests, that will replace the current ones step by step once the new
> tests are reviewed and integrated.

If the tests were based on an outdated version of the spec, then it may 
be better to start from scratch with new tests.

I expect that I will not be able to review tests for the next 4-5 months 
or so. The good news is that I took a lot of notes since 2010 when 
reviewing tests and created a document that you can read. Such document 
lists all kinds of code situations I met when reviewing tests. The 
document is basically how to improve test design, what to avoid, what's 
good to do, what's better, etc.. It's a more advanced test writing 
guidelines kind-of-document.

A first draft was made public in:

28 proposals to improve testcase writing guidelines
http://lists.w3.org/Archives/Public/public-css-testsuite/2011Dec/0004.html

Since then, I have improved such document; it's more detailed and lists 
additional improvements.

Overall, I firmly believe that many tests can be improved in many ways, 
in all sorts of ways. Also, one aspect that soon will have to be 
discussed and agreed on is the necessity to reuse already-created 
reference files (and support files) for tests. We want/prefer an N-to-1 
tests-to-reftest to reduce linkages. As the number of spec and number of 
tests increases, this ought to become more pressing with time.

Gérard
-- 
CSS3 Writing modes section
http://www.gtalbot.org/BrowserBugsSection/CSS3WritingModes/

Received on Monday, 27 October 2014 15:13:00 UTC