dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Wed, Oct 19, 2011 at 09:56:05PM +0200, Anders Logg wrote:
> On Wed, Oct 19, 2011 at 12:51:31PM -0700, Johan Hake wrote:
> > On Wednesday October 19 2011 12:29:51 Anders Logg wrote:
> > > What should we name functions that print stuff? We now have
> > >
> > >   list_linear_solver_methods()
> > >   list_lu_solver_methods()
> > >   list_krylov_solver_methods()
> > >   list_krylov_solver_preconditioners()
> > >   list_timings()
> >
> > I think we should add
> >
> >     list_linear_algebra_backends()
>
> Good suggestion.
>
> > > Is list_foo good? Other options are
> > >
> > >   print_foo()    # Natural, but perhaps a bit vulgar?
> > >   info_foo()     # Goes with the info() function
> > >   display_foo()  # I don't like this one
> >
> > Not a big deal, but I like either one of the later rather than list_foo.
> >
> > What do you have against display or disp?
>
> It sounds more like graphics than text. And we just to have foo.disp()
> in DOLFIN before we had info(foo), so that might also be a reason why
> I don't like it.
>
> > Is the connection to info in info_foo confusing rather than clarifying? If we
> > go with this one, we should probably use the log system (info) to actualy
> > display the info.
>
> We do use the log system (dolfin::cout).
>
> More comments? Let's settle on something "everyone" likes so we can
> get it into the book and 1.0.

More input on this? If not, list_foo is already in the DOLFIN chapter
and unit tests.

--
Anders