On Thu, 2003-11-06 at 21:41, Brett Porter wrote:
> I imagine:
> So I guess type != ext, which makes sense.
It doesn't always and that can certainly be handled by the tools where
you have a handler that deals with a specific type of artifact. It can
choose for the type and ext to be the same or not. Often people like to
classify their ejb's but they are packaged in a JAR with a .jar
extension. The type != ext is easy to deal with.
> I would prefer, for example:
> I prefer this for the same reason that I prefer to have the version
> number in the filename. I think this also follows a well known
> convention that feels natural, even if the filename does get long in
> some cases.
> Continuing to think of it from a tool perspective, the 3 elements of
> type you have here are:
> 1) artifact type (dist)
> 2) release type (src/bin/bin-win32/etc)
> 3) artifact format (extension: tar.gz, tar.bz2 - only makes sense for
> some artifact types)
> > -----Original Message-----
> > From: Noel J. Bergman [mailto:[EMAIL PROTECTED]
> > Sent: Friday, 7 November 2003 1:32 PM
> > To: [EMAIL PROTECTED]
> > Subject: RE: Proposals
> > > > > http://<host>/<group>/jars/<id>[-<version>][-<type>].ext
> > > > Is jars/ to be <platform>/ for the non-Java crowd?
> > > No. In Maven, it's the artifact type, e.g. jar(s), war(s) exe(s).
> > How do we address additional portable and native platforms,
> > e.g., http://www.apache.org/dist/httpd/binaries/? For that
> > matter, what is the mapping for
> > http://www.apache.org/dist/httpd/, which > includes other types
> > of artifacts, including source packages?
> > --- Noel
Jason van Zyl
In short, man creates for himself a new religion of a rational
and technical order to justify his work and to be justified in it.
-- Jacques Ellul, The Technological Society