Re: [Savannah-hackers-public] Keeping web pages in something other than CVS

[Karl Berry]

Hi Glenn,

    In other words, treat CVS just as a means to get the data onto
    www.gnu.org, not as a repository for the history.

Hey, sounds good.

    Maybe tools like this exist already.

I know there is a way to do git-over-svn (never done it myself).
Maybe there is also {git,hg,...}-over-cvs.

    I imagine it just needs a bit of care with added and deleted files.

And symlinks -> .symlinks.

Another small point is that savannah projects have only the one
hardwired slot for the "web pages repository".  So (as a practical
matter) the VCS-X web repo would have to be a separate branch, or
whatever, in the "source code repository" for vcs-x.  I imagine anyone
who wants to go to the trouble of doing this wouldn't have a problem
with that.

    Maybe distribute the scripts via Savannah so that anyone can use
    them (requires a few minutes of Savannah admin time)?

If and when the scripts exist, sure, that is easy.

    Or actually run them on Savannah (requires a bit more Savannah admin time).

Wouldn't want to commit to that (no pun intended :), I can imagine quite
a few complications.  One step at a time, anyway.  First someone needs
to write the things.

    Or maybe such scripts belong more in gnulib than on Savannah, I
    don't know.

More here.  The GNU infrastructure support stuff is pretty tangential to
gnulib's main purpose; gnupload and the like are really only in gnulib
just because it's a convenient place for it.  And by the happenstance of
who wrote it :).

Thanks,
k