dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Mon, Apr 07, 2008 at 06:45:23PM +0200, Anders Logg wrote:
> On Mon, Apr 07, 2008 at 05:30:16PM +0100, Garth N. Wells wrote:
> > 
> > 
> > Garth N. Wells wrote:
> > > 
> > > Ola Skavhaug wrote:
> > >> Anders Logg skrev den 07/04-2008 følgende:
> > >>> On Mon, Apr 07, 2008 at 10:45:51AM +0100, Garth N. Wells wrote:
> > >>>> Anders Logg wrote:
> > >>>>> On Sat, Mar 29, 2008 at 12:48:31PM +0000, Garth N. Wells wrote:
> > >>>>>> The TODO_buildsystem file is almost ready to be removed. Should we 
> > >>>>>> consider a release once TODO_buildsystem can be removed? Are there any 
> > >>>>>> pressing issues which need to be addressed before a release? Does the 
> > >>>>>> build system need more testing?
> > >>>>>>
> > >>>>>> Garth
> > >>>>> I also think we should try to make a release as soon as we can, maybe
> > >>>>> Friday? It should be easy to fix the remaining issues (or at least
> > >>>>> understand what goes wrong and plan for fixing it later).
> > >>>>>
> > >>>> Are things ready for a release?
> > >>>>
> > >>>> Garth
> > >>> Yes, almost. I'd like to get the two SWIG warnings fixed:                                                                    
> > >>>                                                                                                                              
> > >>> 1. The PETScObject thing (should be possible to fix)                                                                         
> > >> I'm not sure this is an easy fix. Non-exposed base-classes are tricky with
> > >> SWIG and warnings. I'll try, though.
> > >> Ola
> > >>
> > > 
> > > I just changed the order of the includes in dolfin_la.h so that 
> > > PETScObject.h appears before any other PETScFoo.h files and the problem 
> > > seems to be fixed.
> > > 
> > > Garth
> > > 
> > > 
> > >>> 2. The warnings for ODE (planning to remove Mmatrix and Jmatrix)                                                             
> > >>>
> > 
> > 
> > Looks like Anders is holding up a release :).
> > 
> > Garth
> 
> It's happened before...
> 
> I just want to make sure there are no major objections to removing the
> functions.

Should we aim at releasing 0.8.0 sometime before the summer?

Basically, it could be 0.8.0 = fully functional linear algebra (both
in C++ and Python).

Then we should be able to do 0.9.0 = fully functional parallel
assembly and solve at the end of the year.

Then next year, we might be able to do 1.0 = everything works,
including the manual.

-- 
Anders