Re: cleanly replacing the trunk (but keeping one branch)

[Jim Hyslop]

Sergei Organov wrote:
> Jim Hyslop <address@hidden> writes:
>
>
> > Jim Hyslop wrote:
> >
> > > Nikita Tovstoles wrote:
> > > [...]
> >
> > > > checked into the repository.  What I'd like to do now,
> > > > is to keep the branch as it is, but completely remove
> > > > the trunk and replace it with all the new code that is
> > > > on my laptop at the moment. I've been the only
> > > > commiter so far. How can I do that?
[...]
> As far as I understood, the OP doesn't want to replace the HEAD with the
> branch, and that's what the above commands do, right? For me it seems
> that the OP wanted to replace the HEAD with his modified files that he
> initially got from the HEAD, so it sounds more like regular commit (with
> a lot of changes/ file renames).

Yes, I think you may be right. I made a bad assumption, that the code on 
the laptop was on the branch.

Nikita, here is a revised sequence of commands that will do what you 
want. This sequence will work regardless of whether your laptop's code 
is checked out from the trunk or from a branch. **NOTE** Any changes you 
may have checked into the trunk from a different location (i.e. a 
different computer, or a different working directory on your laptop) 
will be discarded!

All commands are issued from the base directory of your project, on your 
laptop. Using UNIX-style comments to explain what's going on:

# First, for safety's sake, check in all working code in the
# current directory into a temporary branch. IMPORTANT: you must
# use the 'cvs tag' command, NOT the 'cvs rtag' command.
cvs tag before-major-checkin
cvs tag -b before-major-checkin-branch

# Now, switch to the branch.
cvs update -r before-major-checkin-branch

# at this point, you are now working on a temporary branch.
# Code can be safely checked in, and it will not require
# any updates nor merges.

# 'cvs add' all new files and 'cvs remove' all obsolete files. This step
# is not strictly necessary, but it will ensure that the temporary
# branch is a complete, accurate snapshot of your working directory.
cvs add [as required]
cvs remove [as required]

# Commit the code to the branch. There
# should be no errors or warnings on this commit.
cvs commit

# At this point, all your changes are safely stored in CVS.

# Now, we go back and do the sequence mentioned earlier.
# First, update to the trunk:
cvs update -A

# Now, pull the code off the temporary branch into the current working
# directory:
cvs update -j HEAD -j before-major-checkin-branch

# The following line should produce absolutely no output:
cvs -q di -r before-major-checkin-branch

# Verify that all files are correct. Once you are satisfied that it
# worked properly:
cvs commit

--
Jim