Re: Support for VCS operations

[Yavor Doganov]

Калоян Доганов wrote:
> 
> Let me remind you that we've decided to use an existing infrastructure
> (not only installed Subversion server, but also an existing
> repository)

Yes, I remember this well and I'm neither complaining nor accusing you
or myself.  The rationale was to start quickly and come up with a
prototype, and then think about VCS issues.

However, I've been thinking about CVS difficulties from the very
beginning (remember that mail to Karl about the timestamps -- it is
moot now, I think, after I dived a bit in CVS's internals).

> So, it looks like it's time to switch to CVS.  We can set up one on
> Corellia, or may be we can use the trans-coord repository on Savannah.

I am reluctant to use trans-coord before we decide what would be the
structure, because moving directories in CVS is difficult and you
can't erase a directory.  But probably it's worth the sacrifice.

Savannah is thousand times more reliable, and all leaders already have
write access for the test phase, so no need to bother with forwarding
ports, choosing an authentication mechanism and setting it up.

> I see there is already a `graveyard' module there, may be we can add
> `gnun'?

It's only a directory, not a module.  graveyard is for things that are
dead (or mostly dead, but could be resurrected).  The structure should
be the same as is now and should remain the same even after GNUN is
uploaded to "www".  Of course, that may change if the webmasters say
it should not be /prep/i18n but something else.

Root
  |
  |- gnun
  |     |
  |     |-home.shtml
  |     |-keepingup.html, etc.
  |     |-prep
  |     |    |
  |     etc. |-i18n
  |               |
  |-graveyard     |-i18n.mk
                  |-GNUmakefile
                  |-TODO (no sense to keep it as it is now, IMHO)