Hi Koji-san, I think the current process depends on the size of the change. For small or medium-scale changes, a patch (or even a direct commit) is fine. For something like an entire connector, however, it makes sense to create a branch, since such a substantial patch may be a challenge to get right and to review thoroughly. This is especially true when more than one person is working collaboratively on the same ticket.
As for this point: >>>>>> If no patch files are attched in JIRA, people in latter-day will have hard time to chase the history of improvement/communication cycle on the ticket. <<<<<< SVN does maintain the history of the branch even after it is deleted. So there is an audit trail, which is pretty obvious as long as people note that a branch was indeed created. Karl On Sun, Jun 23, 2013 at 7:35 PM, Koji Sekiguchi <[email protected]> wrote: > MCF Committers, > > There seems to be an accepted practice in MCF community todays says that > making a branch rather than attaching a patch for a new JIRA. > > But I'd like to look into a patch rather than a branch, because if the new > function is acceptted, the diff is going to be merged, then the branch is > removed. > If no patch files are attched in JIRA, people in latter-day will have hard > time > to chase the history of improvement/communication cycle on the ticket. > > I'm positive using braches for development, but I'd like to see patches > for some > checkpoints. > > Thought? > > koji > -- > > http://soleami.com/blog/automatically-acquiring-synonym-knowledge-from-wikipedia.html >
