W3C home > Mailing lists > Public > public-qa-dev@w3.org > February 2003

Re: [check] 0.6.2 release (Was Re: inaccuracies in validator)

From: Ville Skyttä <ville.skytta@iki.fi>
Date: 22 Feb 2003 22:02:33 +0200
To: QA Dev <public-qa-dev@w3.org>
Cc: Terje Bless <link@pobox.com>, Frederic Schutz <schutz@mathgen.ch>
Message-Id: <1045944153.1280.333.camel@bobcat.ods.org>

On Fri, 2003-02-21 at 16:36, Terje Bless wrote:

> >source/index.html should mention the CVS branches, ie. the difference
> >between HEAD and validator-0_6_0-branch.
> 
> Current CVS does. It also contains a link to http://validator.w3.org/dist/;
> a directory that is not version controlled and where packages (tarballs,
> RPMs, and Debs) can be placed. The "Source Code" link in the navigation
> menu now also always points to http://validator.w3.org/source/ instead of
> to the local installation.

Ah, I wasn't aware of that.  Sounds good.

> That leaves just links to your main repositories for RPMs and Debs and some
> explanations for users of same. Could you (Ville, Frederic) send me
> "patches" for whatever you need to make the binary dists work to your
> satisfaction?

My RPM doesn't have a clean patch at the moment, but a hunk of perl code
that does some in-place edits to check.cfg and the httpd.conf snippet. 
It's a perl hunk instead of a patch because it uses stuff provided by
RPM, basically some directory locations which is better from a RPM point
of view than a patch with "hardcoded" paths.  And it avoids the need for
a separate patch file.  You can see the hunk in the spec file in CVS,
look for "# Localize config files".  Yell if you need clarification on
that.

> >The version number of the validator should be somewhere in the UI. Maybe
> >a <meta name="generator" content="W3C Markup Validator 0.6.2" /> would
> >be enough.
> 
> <meta name="revision" content="1.305.2.12" />
> 
> The output contains the CVS revision number of the "check" CGI script (as
> well as the version number of the HTML Template). Do you think we need the
> "human readable" version number also?

Yes, I do, as long as we have the "human readable" ones in Bugzilla and
refer to them in public elsewhere.  Of course, that causes some extra
things for the "release manager", but there are other places where this
is needed anyway.  So it's just one addition to the list.

> >By the way, is it "MarkUp" or "Markup"?
> 
> It's "MarkUp" just because I happened to like StuddlyCaps there that day
> since that's how w3.org tends to spell it. If it irks you, feel free to do
> a global search and replace on it. :-)

The reason why I asked is because I have a feeling that w3.org does
*not* tend to spell it in camelcase :).  See eg.
<http://www.google.com/custom?q=markup&sitesearch=w3.org&domains=w3.org>...

I can do the search and replace, but would like a confirmation which way
is "correct" from a W3C rep.

> That leaves some documentation work from Olivier, and some collaboration on
> the Release Notes (since I've lost track of what's happened since 0.6.1).
> Y'all seem to have fixed all the bugs that cropped up (at least, there's
> nothing in Bugzilla indicating any outstanding issues).

    cvs log -NS -rvalidator-0_6_0-release::validator-0_6_1-release

...is your friend (change the release numbers as appropriate).  I think

    cvs log -NS -rvalidator-0_6_1-release::

...should work too, but I get a "Terminated with fatal signal 11" when
trying it.

> Anyone have anything left to deal with?

Apart from a couple of trivial documentation updates (checklink needs
Config::General), nothing here that should absolutely go into 0.6.2. 
There are some checklink bug reports that I'll be working on soon(tm),
but nothing critical that would prevent this release.

-- 
\/ille Skyttä
ville.skytta at iki.fi
Received on Saturday, 22 February 2003 15:02:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 19 August 2010 18:12:43 GMT