[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: gerris - Gerris Flow Solver

On Thu, 2008-12-18 at 12:30 +1100, Drew Parsons wrote:
> On Wed, 2008-12-17 at 08:17 -0500, Ruben Molina wrote:
> > Hi Drew,
> > 
> > > Ruben, what is the state of your packaging?  Is it ready for upload or
> > > would you be more comfortable if I did some fresh packaging myself?
> > 
> > There is a prospective package on mentors [1]. I have done some tests
> > and it seems to work fine, but I still have some piuparts problems.
> Looks good.  I'll take your work and continue from there.
> > 
> > > I'd also like to register it under the auspices of the Debian Science
> > > team, if everyone's happy with that.  (Got a few packages of my own to
> > 
> > Completely agree. Please feel free to prepare a git/svn repo for gerris
> > on debian-science and we can work together on it :)
> > 

I have created a git repository on alioth in the debian-science area.


I have uploaded Ruben's work there, adding a couple of small changes
(the initial commits are Ruben's work without changes).

I have currently set it up with two branches, debian-unstable and upstream.

I decided one upstream branch should be sufficient, with the various
upstream revisions identified by tags.  I have currently loaded Ruben's
dfsg orig tarball (it has modules removed).   My intention for ongoing
upstream revisions is to go to our upstream branch, pull from
upstream's darcs repo, and tag with upstream/0.9.x+darcsYYYMMDD (so the
current tag is missing a 20 from 2008).  Then any necessary DFSG
manipulations to the upstream source can be made, and retagged as
upstream/0.9.x+darcsYYYMMDD-dfsg.1 (the .1 leaves room for any
subsequent dfsg corrections).  The orig tarball would be generated
directly from this branch (assuming upstream never gets around to
releasing an official stable tarball!)

I labelled our debian branch as debian-unstable rather than just debian
on account of this is how the XSF does it with the Xorg packages.  I
think the idea was that it makes maintenance of stable (etch, lenny
etc) somewhat simpler than pushing tags back and forth.  Any version
destined for experimental would go in a debian-experimental branch.

The code builds fine (including under pbuilder) but when I tested
against my own gerris scripts it failed desperately bad.  I don't know
what the problem is and will worry about it next time when I have time,
but till then I figured it's worth pushing out the git repo so any
others can check or work on it.



Attachment: signature.asc
Description: This is a digitally signed message part

Reply to: