gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Leo Simons <m...@leosimons.com>
Subject Re: New Project for NightlyBuild
Date Fri, 04 Mar 2005 15:07:48 GMT
Hi Sean!

On 02-03-2005 22:02, "Sean Schofield" <sean.schofield@gmail.com> wrote:
> Our project (MyFaces) was just promoted to top level Apache status.
> We're now in the process of upgrading our website and infrastructure.
> We'd like to start doing nightly builds and when I started asking
> around someone recommended using Brutus.
> 
> I've read the Wiki on nightly builds and I'm a little confused.  We're
> happy to supply you with a formal request using the template provided
> but I think we need a little more information on how to go forward.

Quite understandable :-D. Brutus is a machine managed by the gump people for
the purposes of running Gump. There's obviously also a need at apache for
doing nightly builds (as opposed to gump-based builds, which are usually
kinda different). I've been experimenting with a little bit of shell script
to automate that, which is sort-of documented on the wiki and which sort-of
works. We do various experiments on gump, for example the forrest people are
doing stuff there as well. Nothing formal about it :-D.

We're not really providing nightly builds as a "service", nor could we,
since the machine in question (brutus) is very very insecure (because it
runs gump) and nothing it produces should be trusted for any purpose
whatsoever.

I'm also kinda busy doing various other stuff, so if things don't work,
we'll, you're not really getting good "support" at all. For example the
excalibur nightly builds have been broken for months; the only ones
currently working are those for ant. Volunteers welcome :-D

If that's what you want, then...

> I have a few questions such as:
> 
> Does your process handle CVS checkout?  If so do you just checkout the
> buildfile or the whole project?

No checkouts, but it does do 'cvs up'. So I run the checkout manually once
:-D

> Do we need to setup the chron, etc. or do you do all of that once we
> provide you the information in the template.

Ehm, from memory, the latter.

> Where do the resulting jars, etc. get copied to.  You ask for the URL
> to be published to but how do you know which files to publish and
> where they are located?

http://brutus.apache.org/~nightlybuild/builds/$project/

You'll need to specify $project, the rest of that URL is hard-wired into the
script. We publish files that look like distributables:

  files=$(find .                               \
    \( -path '*build*' -or -path '*target*' \) \
    -and \(                                    \
      -name '*.gz'                             \
      -or -name '*.zip'                        \
      -or -name '*.bz2'                        \
      -or -name '*.jar'                        \
    \)                                         \
    \( -not -path '*.svn*' \)                  )
  if [ -n "$files" ]; then
    echo $files >> $log
    cp -f -v --target-directory $logdir $files

> TIA for any help you can provide.

Hope this helps! It isn't exactly a great setup, and isn't intended as such.
In the long-term the plan is to have a Virtual Machine farm on several
machines and different PMCs will get their own VM. But the infra team is
kinda swamped just keeping up with regular admin duties, so it'll be a while
before that happens. If there'd be more volunteers for basic admin, it'd
happen quicker :-D

Cheers,

- LSD



---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@gump.apache.org
For additional commands, e-mail: general-help@gump.apache.org


Mime
View raw message