W3C home > Mailing lists > Public > public-test-infra@w3.org > July to September 2017

Re: Unifying testsuite policy and getting rid of CSS exceptions

From: Philip Jägenstedt <foolip@google.com>
Date: Sat, 23 Sep 2017 14:26:07 +0000
Message-ID: <CAARdPYduQ1aQBgBUEG2Ada=RwykWfaNGMT3aUot2+VYhBO2swg@mail.gmail.com>
To: James Graham <james@hoppipolla.co.uk>, public-test-infra@w3.org
On Sat, Sep 23, 2017 at 9:07 PM James Graham <james@hoppipolla.co.uk> wrote:

> On 9/22/17, 7:31 PM, "Philip Jägenstedt" <foolip@google.com> wrote:
> >
> >      * Keep requiring <link rel=help> for CSS WG stuff.
> >      * Don't require links to specific spec sections if upstreaming new
> tests in bulk, because then it just won't happen.
> I don't entirely follow the difference between this point and the
> previous one. Is the difference "in bulk" — in which case how do you
> make the lint work, or "specific spec sections", in which case what's
> the point compared to just ensuring that the tests are in the
> spec-appropriate directory?

This wouldn't be lint enforced. The lint wouldn't require a section link,
but reviewers would expect to see it and could ask about it. But if a large
number of tests are being upstreamed, it's fine to just link to the spec,
and if a reviewer doesn't like it they get to do the work themselves after
the fact.

The point would be that spec links are versioned, or at least could be made
versioned when bumping the spec level. The implementation repot generator
would use this number instead of the number in the directory name, so we
can have one directory.

I asked Emil Eklund about my guesses about pain points, and heard back that
"I'd argue that the versioned/unversioned and duplication concerns are the
worst and would gladly help to clean that up."
Received on Saturday, 23 September 2017 14:26:41 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:34:13 UTC