[GRASS-dev] cvs2svn questions

[Martin Landa]

Hi,

2007/10/25, Glynn Clements <glynn at gclements.plus.com>:
> Hamish wrote:
>
> > there are a few questions listed here:
> >   http://grass.gdf-hannover.de/wiki/Migration_from_CVS_to_SVN
> >
> > to address two of them:
> >
> > 1. remove deleted files in the Attic?
> >
> > my suggestion would be keep them, unless they cause a big problem and
> > waste huge amounts of space. Sometimes there is interesting stuff
> > there or history as to why it changed.
> >
> > Also, is it possible to check out older versions of the code from
> > before the file was moved there if you delete the attic??
>
> No. But (AFAICT) there's no straightforward way of getting those files
> into a Subversion repository.
>
> Most of the files which were in the attic are there for the same
> reason.
>
> Originally, 5.7 was an "overlay" on top of 5.3 (the "make mix" era).
> When a file needed to be changed for 5.7, the modified version was
> added to the 5.7 repository.
>
> Sometimes, the changes were backported to 5.3, or 5.3 was updated in a
> way which meant that 5.7 no longer needed a separate version of the
> file, so the file was removed from the 5.7 repository. When "make mix"
> was abandoned and the outstanding files were copied into the 5.7
> repository, the earlier versions end up in the attic.
>
> So, most of the files in the attic are versions from the early days of
> 5.7 which diverged temporarily from the 5.3 versions (files which
> diverged permanently will still be there, rather than the 5.3
> versions).
>
> Those files are of no use to anyone other than an archeologist digging
> up the prehistory of 6.x. IMHO, it isn't worth the effort of figuring
> out how to preserve them.

ok, then the Attic files can be removed, I guess. Take a look e.g. at

http://josef.fsv.cvut.cz/cgi-bin/viewcvs.cgi/grass6/display/d.erase/Attic/?root=grass

exactly this case.

> > 2. creation grass7 repo.
> >
> > the suggestion in the wiki page is to run cvs2svn for grass6 twice.
> > why not copy it after? svn must have a way to clone a new repo from an
> > old one. (but whatever is easier & cleaner)

First the grass6 svn is created (trunk, selected branches and tags)
based on grass6 cvs, then grass7 svn based on grass6 HEAD (*only
trunk*, no branches, tags).

Martin

[snip]

-- 
Martin Landa <landa.martin at gmail.com> * http://gama.fsv.cvut.cz/~landa *