[plt-scheme] General protocol for breaking the trunk :-)

On Jul 20, Noel Welsh wrote:
> If I commit the changed uri-codec to the trunk I will break
> many things -- at least the web-server, help-desk.  So,
> should I work on a separate branch?  This branch will exist
> for a long time, as I don't have the time to fix the
> web-server and help-desk in a short period.  So this branch
> may get out of sync with the rest of the tree (e.g. Jay's
> changes).  So perhaps branching isn't a good idea. 
> Suggestions?  The only alternative I can see is we create
> one branch (called, say, bytes) for all the work shifting
> the web infrastructure to use bytes.  Then we can hack away
> at a leisurely pace, and not worry about breakage or
> merges.

[Sorry for the late reply, my mailbox is a frightening thing to look
at.]

Branching should be fine, even for long term.  First, there's the
obvious thing -- if you don't step on anyone's toes, then it would
just work as you expect it.  If you do require changes to the web
server then you can do that on the branch too.  If you need to work on
things together with Jay, then things are even better.

If you do want a general branch and be able to merge things
occasionally from the trunk, and eventually merge all your stuff back
on the trunk, then that's not a problem too.  All you need to do is to
merge changes on the trunk onto your branch, and just keep the revision
number somewhere (preferably in the log message) so you know where to
do the next merge and what commits to your branch to not merge back on
the trunk when the time comes.  There is an alternative way, where at
each merge-from-the-trunk point you restart your branch from scratch,
I describe that in the internal.plt-scheme.org/svn.html page.

-- 
          ((lambda (x) (x x)) (lambda (x) (x x)))          Eli Barzilay:
                  http://www.barzilay.org/                 Maze is Life!