[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Move to git?

Just a few brief notes rather than going back and forth in arguments.  I may respond with details on what I am thinking in responses to posts later but I don't want to get dragged down into those.  Rather I want to discuss current considerations.

First, I think that everyone would agree that good support for git users is important for the project.  Having a viable git ecosystem means that consultants can mix and match features more easily than they can if everything is on svn.  This is a powerful thing.  Having good support for it is good.

However, with flexibility comes complexity, and I doubt that anyone can deny that either.  I think that since currently we are not limited by svn's model (we could be in the future though) it would be a mistake to move the project there.  I think we have to assume going in that a lot of things would have to be thought out, decided, etc. and even if we committed to a move today, who knows when it would really happen?  There is a lot to be said for letting things organically develop and keeping open the possibility of moving when it becomes a win for the project's development to do so.

There are certain things that could happen that would make it a clear win.  For example if we ended up with mini-forks aimed at solving problems for certain market sectors (vertical or horizontal) then I think it would probably be a win at that point.  However, switching for the sake of switching strikes me as a bit of a solution in search of a problem.

One other major point in favor of allowing things to organically develop is that it gives time for some issues to be reviewed and settle in.  As problems arise we get a chance to look at them more carefully and the solutions we use if/when moving to a distributed version control system will likely build on those.  This means waiting can actually reduce the complexity of a move if instead of moving today we are enabling people to use other tools today.

So with those thoughts in mind I want to ask two basic questions:

1)  For git users, what do we need to do today to make use of git with LedgerSMB more pleasant?

2)  For those who currently use git what sorts of processes can we have today, without moving currently, that will help you get more out of using git with ledgersmb?  This is an important one because even if we move in the future, it isn't clear we'd entertain direct pull requests.  We might use git just to make it easier for people to maintain mini-forks and share features between such.  (My understanding is this is largely how PostgreSQL uses it for main development, and my guess is that virtually all patches still come in via email.)

Best Wishes,
Chris Travers