Re: [Savannah-hackers] submission of Savannah - Next Generation - savannah.gnu.org

[Elfyn McBratney]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wednesday 12 May 2004 11:34, Sylvain Beucler wrote:
> Elfyn McBratney wrote:
> > I've submitted this project so we have a place to store our changes
> > to the Savane code, new features, GForge stuff (for when we migrate),
> > Sylvain's bric-a-brac docs, &c in a single location.  I was going to
> > branch out the savannah repo for this, but I thought It'd be nice to
> > preserve it's history :)
> >
> > OK to approve?  Yay?  Nay?
>
> Since nobody gave any valid reason to move to GForge, I am still
> opposed to that. I am still waiting an answer from Bradley and the
> sysadmins to an e-mail I sent twice at the end of the migration/merging
> discussion.

As am I.  But it seems like the points we raised do not matter as there have 
been no answers (apart from RMS' emails toward the end).  Perhaps it's time 
to raise those points again?

I had intended to send an e-mail out to everyone to try and arrange a 
discussion between all of us (savannah-hackers, the FSF) regarding the 
migration but haven't had the time..

> My bric-à-brac (plus "hacking savannah.gnu.org") should be kept private
> for now because I am not sure at what point some information should be
> kept private, hence why the CVS repos is at svadmin and not in the
> savannah project.

Fair enough :)

> I planned to make a branch in the savannah project, from DEV_2003-09-
> 05_CERN, using a DEV_2004-05-XX_GNU name. CVS already allows us to keep
> the history of the project.
> The goal is to find each non-commited change in our installation and
> commit them to the branch. Before, we have to move each 'cvs co'
> present in the system to the new branch. Then we have to make our
> installation clean and well configured, so that we can 'make install'
> to apply our changes - I think it is the cleanest way to update all the
> components of the installation, which are in several places and with
> some changes from the source form.

This was my first thought (to branch), but when I tried to tag (a couple of 
weeks ago) it failed with an "access denied" error.  From a brief look, it 
looked like a permissions error in the repo, but I haven't had the time to 
look into it (and forgot to tell you about it ;)

I'm not fussed which ever way we go; we just need /a/ repo where we have a 
working codebase that we can hack on.  There are changes all over the place, 
whose history is live in the filesystem, which leaves us without a backup 
plan if a piece of code gets broken.

I've also been working on code today that parses cvswrappers, 
{commit,log}info, and modules CVS admin files (as part of the CVSROOT/ mgmt. 
front-end), which I would like to get versioned some time soon

> I am planning to do that on my test install but I lacked the time to
> finish it. Changing the branch used by a working copy containing non-
> commited changes may be a bit tricky, so we have to be sure of what we
> do.

I have a tarball of the infamous CERN branch merged with all the changes made 
to the Savane code (perl backend library/scripts, frontend, content/config, 
etc) as of 2004/05/06 which I'm using for my test install.  It's working much 
the same as Savannah is now (same bugs, same features ;)

It should be as easy as `cvs rtag', untar, `cvs add' and `cvs commit'.

But my test install has only two projects, so it is no where near the size of 
Savannah, and probably not a good testing platform for this.  Perhaps we 
should try and clone parts of /savannah and work on that as our test install?  
That way we are working on with 'live' data and tests should be more 
accurate.

> As far as I am concerned, I think it is selfish from us to accept all
> kind of projects only for the FSF (they are displayed on the home
> page), so we should not approve any project unless for testing until
> everything is ok.

OK.  Like I said above, I just need a place where I can start commiting code.  
A new project was my first thought, but after thinking about it branching the 
existing `savannah' project would probably be best. :)

BTW, sorry for my lagging responses.  I'm now bound to office hours (not 
working remotely anymore) so I can only get down to Savannah stuff after 6PM 
GMT 8)

Elfyn

- -- 
Elfyn McBratney, EMCB
mailto:address@hidden
http://www.emcb.co.uk/

PGP Key ID: 0x456548B4
PGP Key Fingerprint:
  29D5 91BB 8748 7CC9 650F 31FE 6888 0C2A 4565 48B4
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAol8CaIgMKkVlSLQRAqtLAJ4tZT2sja6hw5arEAWlNxPaqMmcEQCdHBWe
7xwI//r3nK8a+i6LQVk6oe8=
=Gtpq
-----END PGP SIGNATURE-----