Subject: Re: [boost] Volunteers needed to run regression test reporting
From: beet (r.berlich_at_[hidden])
Date: 2014-01-18 09:38:08
Am 17.01.14 17:56, schrieb Beman Dawes:
> We have had a long term problem with regression test reporting going
> offline unexpectedly. The process used XSLT processing was hard to
> configure, horribly slow, and very fragile, and those factors made it a
> pain to run.
> Last year, Steven Watanabe rewrote the report generation in C++, and those
> issues have pretty much disappeared, so the process is now far easier to
> run. It runs in five minutes or less.
> Hooray C++! Hooray Steven! And Hooray Tom Kent who runs the reports!
> The remaining reliability factors are local ones like system crashes. Since
> the reports run as cron jobs, it can be quite a while before anyone is
> inconvenienced enough to report it to the Boost list, and Murphy's Law
> ensures that it is always the day after the person running the reports left
> for a week-long back country hike.
> We can mitigate that by having several people running the reports. I've
> written some docs, so it shouldn't be too hard, although right now it does
> require a *nix-like system. I've been running them on a Windows host using
> a VirtualBox virtual Ubuntu system, and have found it to be easy.
> One possibility would be for Tom to continue to run the reports on the
> hour, a second person to run them twenty minutes after the hour, and a
> third to run forty minutes after the hour.
> Please volunteer to help make regression reporting more timely and reliable!
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Any chance for the test environment to be made available as an RPM and
Debian package on the Linux side ? I believe you would find a rather
large amount of testers willing to contribute ressources (e.g. at
Universities), if all they have to do is an "apt-get install" of the
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk