W3C home > Mailing lists > Public > public-css-testsuite@w3.org > October 2010

Re: Corrections to the Implemention Reports

From: Linss, Peter <peter.linss@hp.com>
Date: Thu, 28 Oct 2010 20:06:41 +0000
To: Boris Zbarsky <bzbarsky@MIT.EDU>
CC: Simon Fraser <smfr@me.com>, fantasai <fantasai.lists@inkedblade.net>, Tab Atkins Jr. <jackalmage@gmail.com>, John Jansen <John.Jansen@microsoft.com>, "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>
Message-ID: <D8A26AB8-A482-46FB-BCC8-EDB9792DA751@hp.com>

On Oct 28, 2010, at 8:47 AM, Boris Zbarsky wrote:

On 10/28/10 11:37 AM, Simon Fraser wrote:
What I recall from CSS-WG minutes is that the implementation report can be generated with a public beta or nightly build, but that build has to have been available to the public for a minimum of one month.

I have to be missing something.  If I push a change to Gecko, we
generate a nightly with that change, we discover the change breaks sites
and I back it out, then I wait a month I can use that known-broken build
to generate an implementation report?

To be clear, that's not allowed. The exact phrasing is:

A user agent that:
 implements the specification.
 is available to the general public. The implementation may be a shipping product or other publicly available version (i.e. beta version, preview release, or "nightly build"). Non-shipping product releases must have implemented the feature(s) for a period of at least one month in order to demonstrate stability.
 is not experimental (i.e. a version specifically designed to pass the test suite and not intended for normal usage going forward.)"

So, "it was in one build a month ago" doesn't qualify, "it's been in all the builds for the past 30 days" does.

I can _almost_ see the point of allowing betas, since that allows us to
try to generate implementation reports now that will hopefully match
shipping browsers by the time the test suite is stabilized.  Even that's
a bit fishy, given the amount of churn I've seen between beta and final.
 But the nightly thing just doesn't make any sense, unless I'm missing

The thing that really matters is: "is this spec implementable", not "is it implemented now" (of course the proof of implementability is an implementation...). So having nightlies is good enough so long as it's a real effort towards an implementation and not someone trying to game the system.

Received on Thursday, 28 October 2010 20:08:24 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:13:21 UTC