on Mon Mar 19 2012, Olaf van der Spek <ml-AT-vdspek.org> wrote:
I'd be happy to help out.
Is there a concrete plan?
Not a very concrete one. Someone needs to do some planning, too. IMHO, we should set up Redmine at OSUOSL and make a transition to that almost immediately, because the conversion can be pretty much automatic.
Has the decision to move to Redmine been made (already)? Last time some were arguing Trac might be 'fine'.
We've also heard from people who say that they've never used a large Trac installation that wasn't slow. I guess the steering committee hasn't voted on this. If you want a formal decision before proceeding, please request that Beman take a vote in the steering committee.
That will require some work on the Redmine end, I expect. OSUOSL can probably help us with some of the Redmine database rewrites that will be required, but I am not sure of that.
At that point we should probably also give each library its own Redmine project, or, if we are modularized at that point, possibly give library maintainers the option to use the GitHub issue tracker associated with the library's repository there.
Why not move to a GH tracker right away?
The main reasons: 1. We don't know that it's possible to preserve all the information from Trac in such a conversion. 2. GH trackers are not well-suited to multi-project repositories. Until we are modularized IMO it doesn't make sense. 3. Redmine (with the appropriate plugin) supports code review, whereas GitHub only supports change review. That's a significant and problematic difference where the formal review process is concerned. -- Dave Abrahams BoostPro Computing http://www.boostpro.com