Hey Garry,
At Last.fm we have an Ivy repository in Subversion which contains
hundreds of third party libraries that our projects depend on. It was a
bit of work setting this up, but this was split among the development
team who added libraries as they needed them. This initial overhead was
well worth the productivity gains we see now with most libraries already
there. What definitely helped was having a well-communicated, logical,
standard layout of organisation/module/version that is enforced across
all artifacts added to the repository.
We also have many projects which publish various artifacts (jar, war,
tgz etc.) to the same Subversion repository and these then get used by
other systems (which retrieve the artifacts either using Ivy or "svn
export" commands). This repository has been in heavy use daily for over
a year now and has served us extremely well.
<shameless plug>
If you are thinking of using Subversion to store your Ivy repository, we
use IvySvn:
http://code.google.com/p/ivysvn/
(we wrote most of it ;) ) There are some known issues relating to web
dav and repository roots with nested folders so give it a go before
committing to anything (and please report any issues you experience to
the IvySvn project so we can look at them).
</shameless plug>
Good luck!
Regards,
Adrian
Garry Smith wrote:
Hi,
Apologies for the newbie question.
I want to set up an enterprise repository that the developers on my
project can submit to using the publish ivy task, from their
development machines.
The 'repository machine' has apache 2 and also Subversion (behind
apache 2)
The instructions at
http://ant.apache.org/ivy/history/latest-milestone/tutorial/build-repository.html
concentrate on creating the repository locally but not about access.
Can you point me to further information?
The question has popped up a few times in the archive, but I didn't
see a definitive answer.
I was thinking to host the ivy repo via SVN. Is this sensible? Any
gotchas?
thanks in advance
Garry
