W3C home > Mailing lists > Public > www-style@w3.org > September 2016

Re: PROPOSED RESOLUTION: merge csswg-test into web-platform-tests

From: Geoffrey Sneddon <me@gsnedders.com>
Date: Thu, 29 Sep 2016 16:58:10 +0100
Message-ID: <CAHKdfMg3k3993kKm5YhWNQqN8UnxsTOP+z2jNzV2mLQF1SuZ9g@mail.gmail.com>
To: Gérard Talbot <css21testsuite@gtalbot.org>
Cc: Public www-style mailing list <www-style@w3.org>
On Tue, Sep 27, 2016 at 11:44 PM, Gérard Talbot
<css21testsuite@gtalbot.org> wrote:
> Le 2016-09-27 12:23, Geoffrey Sneddon a écrit :
>> 3. Ensure web-platform-tests's documentation is up-to-date and
>> cohesive, both for submitting tests and reviewing them.
> Geoffrey,
> My main request regarding reviewing tests is that reviewing tasks
> (procedures, steps, requirements, areas to check, what's decisive,
> important, critical, etc) will be specified and explicit: what should be
> reviewed, how it should be reviewed, what test authors should expect, what
> is expected from reviewers, from a review, etc... would be specified and
> explicit. The current documentation is, in my opinion, already doing an
> excellent job in that regard as it covered a wide range of possible
> situations and parameters.

Our current documentation seems to be primarily
<http://testthewebforward.org/docs/review-checklist.html>, given the
old documentation at <https://wiki.csswg.org/test/review> has a big
notice at the top. Is there some other documentation I've missed

>> 4. Make https://hg.csswg.org/test/ and http://test.csswg.org/shepherd/
>> read-only. (Really this can be any step up until this point; exact
>> timing doesn't matter.)
> I am sorry... I am not sure I understand what you mean by make those
> read-only (and what would that imply) and I use and have used exclusively
> Mercurial and Shepherd in the last 5 years.
> Eg
> Say I want to add a comment to 001 test in Shepherd:
> http://test.csswg.org/shepherd/testcase/001/
> Will I be able to if http://test.csswg.org/shepherd/ becomes read-only?

See below.

> Eg
> Is https://hg.csswg.org/test/ not already read-only?

It's not, because it's where you push to (i.e., `hg push` uses HTTP
PUSH and PUT requests to that site). More below…

So, nobody seems to have sent any email about what was resolved at the
SF F2F, beyond that in the minutes. I presume everyone took it to be
someone else's problem. :)

The intent is that everything goes through GitHub in future
(especially once merged with web-platform-tests!), to unify where all
our issues are tracked (for similar reasons as to why all technical
feedback on drafts has moved to GitHub; we currently have issues
tracked in Shepherd, on public-css-testsuite, and on GitHub, which
means if you want to find known issues in a given test you have to
look in three places), and to avoid the somewhat awkward
two-way-mirror between Mercurial and Git. Moving to GitHub also allows
us to do review-then-commit, as opposed to our current status quo of
review-then-commit on GitHub and commit-then-review on Mercurial.

Part of this is being driven by the desire to merge everything with
web-platform-tests so that we have all tests for the web platform in
one repository with one way of getting a list of tests to run, one set
of ways of running the tests, and one way to contribute tests. As it
is, web-platform-tests gets far more attention from browser developers
(both in terms of running tests and submitting them), which is
something we rather want!

Essentially, we're asking those currently using Mercurial and/or
Shepherd to move to GitHub. There's some documentation when it comes
to GitHub at <http://testthewebforward.org/docs/github-101.html>, but
I think nowadays there's better documentation elsewhere.
<http://www.wikivs.com/wiki/Git_vs_Mercurial> has a decent comparison
between git and hg, including corresponding commands. (There's a few
things to note: most obviously that git has a staging area between the
working copy and the commit tree, whereas hg does not; as a result,
you have to `git add` files that are already in the tree to get them
in the staging area to then be committed.) I don't know if the Pro Git
book (<https://git-scm.com/book/en/v2>) would be helpful; it may be a
bit too in-depth for many trying to just start using it!

FWIW, there is some plan to extract all the data currently in Shepherd
into GitHub issues, but there's not really currently a timeline for

Received on Thursday, 29 September 2016 15:58:39 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:09:04 UTC