dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Tue, Apr 21, 2009 at 11:32:53PM +0200, Johan Hake wrote:
> On Tuesday 21 April 2009 22:18:45 Anders Logg wrote:
> > On Tue, Apr 21, 2009 at 07:40:14PM +0200, Johan Hake wrote:
> > > On Tuesday 21 April 2009 09:26:52 Anders Logg wrote:
> > > > After removing the FFC form language from FFC (replaced by UFL), the
> > > > DOLFIN Python interface is now broken.
> > > >
> > > > Edits are needed in quite a few places:
> > > >
> > > > site-packages/dolfin/assemble.py
> > > > site-packages/dolfin/errornorm.py
> > > > site-packages/dolfin/form.py
> > > > site-packages/dolfin/function.py
> > > > site-packages/dolfin/functionspace.py
> > > > site-packages/dolfin/__init__.py
> > > > site-packages/dolfin/norm.py
> > > > site-packages/dolfin/project.py
> > > > site-packages/dolfin/specialfunctions.py
> > > >
> > > > I can do some of these, but the most intricate dependencies are in
> > > > function.py and functionspace.py. Mr Hake?
> > >
> > > Please have a look in
> > >
> > >   site-packages/dolfin/ufl
> > >
> > > and you will find some good suggestions on what you need to change. You
> > > might want to
> > >
> > >   rm site-packages/dolfin/ufl/__init__.py site-packages/dolfin/ufl/cpp.py
> > >   cp site-packages/dolfin/ufl/* site-packages/dolfin/
> > >
> > > and then
> > >
> > >   hg rm packages/dolfin/ufl
> > >
> > > there are still some rough edges wrt split, and the allready discussed
> > > TensorFunctionSpace. There are some design descisions in the DOLFIN
> > > library and ufl that might make the split quite difficult, as it is now.
> > > I have not have any time to look more thorough into it.
> >
> > Cool. I wasn't even aware of the ufl subdirectory. 
> 
> Well, I have told you both on the list and in person ;)

Shows you that either I don't listen or I have a very short memory.
It's probably a combination of the two. ;-)

> > When did you add this?
> 
> 24/3. 
> 
> > I'll push these changes and continue tomorrow on making it work.
> 
> Good! 

It seemed to work as a first approximation but there are some minor
things to work out (and I'm not even thinking about the split() issue
yet).

I'll take a look later if I can and if no one else has fixed it
before.

-- 
Anders

Attachment: signature.asc
Description: Digital signature