I have recently migrated a CVS repository to SVN. Now this was the first
pass at the solution and I am not sure that I am doing it the right way
or the suggested layout to make support/branching and working with the
system easier. So I have decided to ask the enlightened people on this
list for some wisdom.
Current Layout:
CVS (Around 60 Eclipse projects):
Project1
Project2
Project3
SVN (Migrated: Approach One)
Toolkit/trunk/base/Project1
Toolkit/branches/0-0-1/base/Project1
Toolkit/branches/0-0-2/base/Project1
Toolkit/trunk/extensions/Project2
Toolkit/branches/0-0-1/extensions/Project2
Toolkit/branches/0-0-2/extensions/Project2
Customer/trunk/Project3
Note:
1/ The Customer set of projects don't have any branches/tags at the
moment
2/ The Toolkit Projects are desired to be in two separate groups but the
commits/version numbers are the same due to the close coupling between
the projects.
3/ The major aim of the migration is the use of atomic commits across
the various projects as we have been having problems with partial CVS
commits across the projects colliding.
Now each of the projects in the SVN are checked out to appear the same
as the initial CVS projects when working in Eclipse. Currently we are
having problems merging and seem to lose files (I think half of this is
probably education of the developers using the subversive plug-in).
What I am looking at is a major refactor of the repository to make the
conversion into Subversion easier to maintain in the long run. Because
currently I am feeling like the current solution is not going to be able
to be maintained for a long time as projects come and go during the
course of normal development.
Thanks
-Andrew
Received on Wed Dec 20 15:41:29 2006