[Python-Dev] Bug day outcome

[Tim Peters]

[Anthony Baxter]
> We have the equipment (xs4all has donated 3 machines) and we've been
> assured on multiple occasions that we can't even make a dint in their
> bandwidth, so that's not a problem.
>
> The question is simply whether we'd be better running our own CVS (or
> SVN) server.

Note that xs4all will not allow running pserver (anonymous) CVS.  This is
for the same reason pserver access is disabled at zope.org now too:  the
history of "security bugs" associated with pserver CVS is dismal.

> The major reason to do this is that the SF anon CVS is awful.  Truly
> awful.

It's a hell of a lot more useful than Zope's anon CVS now <0.5 wink>.

> This impacts people who'd like to contribute, but can't get current
> CVS. A while ago, I suggested taking the tarball of the CVS root and
> putting it somewhere, then setting up 'anoncvs.python.org' and pointing
> it at this. Is this still a good idea? If so, I can look at it.

I'm not sure what this suggestion is.  Daily tarballs of HEAD, and the 22
and 23 branches, are already available here:

    http://www.python.org/dev/

> As far as moving CVS/SVN, if it happens before 2.4(final) I will be very
> cranky <wink>

Good call!  Zope's conversion to SVN has been quite bumpy, and there still
isn't a released version of ViewCVS that works with SVN.  The "download a
tarball" feature of in-development ViewCVS-with-SVN was disabled (at Zope)
when Jim discovered it sucked up 400MB of RAM to generate a 3MB Zope3
tarball.  The Berkeley database often wants an admin to run recovery on it,
and etc:  Martin is correct that the major barrier here is sufficient
volunteers to do support and maintenance.  Source control systems, and
trackers, require endless care and feeding.