Re: Original commit history for gfortran

[Janus Weil <janus at gcc dot gnu dot org>]

>>> We're in the process of considering contributing to gfortran for a
>>> special project, but when we started to vet the codebase we hit a bump
>>> in lack of commit history.
>>
>> what exactly do you plan to contribute?
>
> Let me turn the question - If a handful of people familiar with Fortran
> wanted to contribute in a significant and meaningful way.. What's left? ?(I
> assume bugs as usual and F2K3/F2K8?)

Whoa, that's a big question. The short answer is: A lot!

The long answer goes like this:

First, bugfixing is of course always useful (bugzilla currently lists
about 500 open Fortran bugs, enough for everyone ;).

Second, there are a few major F03 features, which are not implemented
yet (cf. http://gcc.gnu.org/wiki/Fortran2003Status):
 * IEEE arithmetic
 * parametrized derived types
 * finalization
 * polymorphic arrays (I'm not sure if Paul is currently working on this?)
These are the major ones I think, but there are definitely several
smaller things which need to be implemented/finished/improved.

Then there is F08, where we also miss major features (cf.
http://gcc.gnu.org/wiki/Fortran2008Status), most prominently:
 * submodules
 * coarrays (which Tobias Burnus and Daniel Carrera are working on)
and several smaller items.

In general we are also urgently in need of an array descriptor update,
which is useful e.g. for the maximum rank of F08, CLASS arrays,
coarrays etc. (cf. http://gcc.gnu.org/wiki/ArrayDescriptorUpdate).
Furthermore, performance or diagnostic improvements can not hurt
either.

Any contributions are highly welcome.

Cheers,
Janus