Re: Proposed gfortran development branch

[dominiq at lps dot ens dot fr (Dominique Dhumieres)]

> All who agree, please reply "yea", those opposed, "nay".

I am preparing a more argumented reply, but my binary answer is
"nay". First I share the Janne and Richard's opinion about the
possible difficulties associated with a branch. Second I think 
the gfortran maintainers should not follow the "belt and suspenders"
strategy. It made some sense when the release was "a week" away.
This is no longer the case and since the I/O performance is marked
as a regression, the corresponding reviewed patches should be commited
and the others should be reviewed and commited. Also the patches for
wrong codes should be processed.

This will give a much better test coverage than a branch that very
few people will use (for instance I am not in a position to look at
mre than the trunk on darwin).

Dominique