The simplest way to convert/import legacy CVS repositories to GIT

Another task that should be relatively simple seems to be unnecessary complicated everywhere - so a quick note here.

Assuming that you have the CVSROOT (the tree with CVS repositories to convert) locally available (a network drive should be OK, as long as you can access it directly using cd, ls etc.), all you have to do for each cvs repo (module) is:

$ mkdir module_name ; cd module_name
$ git cvsimport -v -d :local:/home/repo/cvs module_name

And you have the cvs module in a new git repo, you can eg. add a remote and push all there, eg. for github:

$ git remote add origin git@github.com:a_user/a_repo.git
$ git push -u origin --all
$ git push -u origin --tags

A bit more complicated way…

… using cvs2git (cvs2svn) is described on this page.

The shortest way:

$ mkdir repo_converted_to_git
$ cd repo_converted_to_git
$ git init
$ cvs2git \
    --blobfile=cvs2git-tmp/git-blob.dat \
    --dumpfile=cvs2git-tmp/git-dump.dat \
    --username=cvs2git \
    /path/to/cvs/repo
$ cat cvs2git-tmp/git-*.dat | git fast-import

In case you need more detailed settings (like properly converted information about the commiters) it might be better to use the cvs2git with an options file:

$ cvs2git --options cvs2git.options

The most important in the cvs2git.options is the path to CVS repo:

run_options.set_project(
    r'/home/repo/cvs/cvs_repo_to_convert',
    (...)

Unfortunately, it cannot be set in command-line - pity, as the tool would be much more flexible in use, maybe one day the developers of cvs2git/svn will change this…

The options file also contains paths to the generateg git blob/dump files:

ctx.tmpdir = r'cvs2svn-tmp'
ctx.revision_collector = GitRevisionCollector(
    'cvs2svn-tmp/git-blob.dat',

and setting to properly transform information about commiters. eg.:

author_transforms={
    'jrandom' : ('J. Random', 'jrandom@example.com'),
    'prnd'    : 'P. Rnd <prnd@example.com>'
}

The archive/package with cvs2git (cvs2svn) should contain an example of such file which is a good starting point.

A bit more complicated way…

Useful related links