[Top][All Lists]

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

Re: CVS queue tool?

From: Greg A. Woods
Subject: Re: CVS queue tool?
Date: Wed, 3 Jul 2002 16:36:02 -0400 (EDT)

[ On Wednesday, July 3, 2002 at 15:28:24 (-0400), Derrek Leute wrote: ]
> Subject: CVS queue tool?
> Is there a tool that will let me make a queue of cvs commits while I am
> unable to connect to the repository? The goal is to take my laptop on
> trips with an up to date copy of my projects. I don't want to make one
> large commit with one simple message. I prefer to make many smaller
> commits with messages specifying exactly what has changed and why.

Yes, it's called your working directory and a shell script.

You can't partially commit anything -- there's no meaning to doing so.
Just keep your changes in your working directory and commit them when
you connect.

If you plan on making multiple changes that you wish to separately
commit then the safest thing to do is to clone your unmodified working
directory and make each change in its own separate working directory.

If some of your changes depend on others and you wish to commit them
separately then you should connect to the repository and commit them
as each node on the dependency graph is reached, especially if there are
other people concurrently working on the same files.  It's much safer
and easier to do merges in small steps if you want to keep the change
sets between commits usable.

You could clone your working directory between each change and then
you'll have a series of working directories that must be committed in
the proper sequence, though this will be a bit messy as you'll have to
resolve confilcts within each from the previous as you go along
(remember you have to be "cvs update" before you "cvs commit").

Your other option is to create a set of patch files (i.e. diff output)
and then apply those to a working directory, committing between
applictions, once you're connected to the network again.  This way you
can make a cumulative set of changes that follow a dependency chain, and
you'll have to resovle conflicts before you commit each patch set.

> I *could* run a mirrored repository on my laptop and then merge that with
> the primary repository when I return.

Not unless you have some magical repository merge tool.

(I've seen people do this with SCCS files, but never RCS files, and it
would be doubly dangerous with CVS)

                                                                Greg A. Woods

+1 416 218-0098;            <address@hidden>;           <address@hidden>
Planix, Inc. <address@hidden>; VE3TCP; Secrets of the Weird <address@hidden>

reply via email to

[Prev in Thread] Current Thread [Next in Thread]