Tracking local port hacks?

[Rob Ellis]

On Tue, Jan 27, 2004 at 04:06:44PM -0600, Dan Nelson wrote:
> In the last episode (Jan 27), Erik Trulsson said:
> > On Tue, Jan 27, 2004 at 02:05:16PM -0500, Rob Ellis wrote:
> > > We sometimes find it necessary to make some small change to a port
> > > before installing it, and need a way to track/merge these changes
> > > as ports are updated. Is there a recommended way of doing that?
> > > 
> > > The cvsup faq (http://www.cvsup.org) suggests that it's possible to
> > > get sources in "cvs mode" and it has some suggestions for managing
> > > a local branch. Is that the best way to do it? Anyone know how BIG
> > > the ports tree is if we get it via cvsup in "cvs mode"?
> > 
> > Using cvsup to maintain a copy of the whole (or part of the) CVS
> > repository, and then using cvs to check out the branch/version you
> > want is certainly one way to do it.  Cvs (unlike cvsup) understands
> > local changes and can merge changed files (assuming the changes don't
> > conflict of course, then you have to do some editing by hand.)
> > 
> > I don't know if it is the *best* way of doing it, but it is the way I
> > do it and it works fairly well.
> 
> I do it too. It takes a bit longer to update, since you have to "cvsup"
> then cvs "update -dP", but I do it at night via cron, and get an email
> the next morning if there were any conflicts with local changes.
> 

Thanks. So you don't have a local branch, just leave your
changes as unchecked-in modifications in your ports 'sandbox' ?

Any thoughts about trying to maintain a local branch? I was thinking
(after more reading about cvs) something like:

   # once, in the sandbox
   cvs tag -b myports
   cvs update -r myports

   # nightly, in cvs mode...
   cvsup

   # after cvsup, in the sandbox
   cvs update -dP -j HEAD

   # after update, check for conflicts, then...
   cvs commit -m 'cvsup changes'

??

- Rob