This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Implementing BLOCK
- From: Craig Powers <enigma at hal-pc dot org>
- To: Fortran List <fortran at gcc dot gnu dot org>
- Cc: Daniel Kraft <d at domob dot eu>
- Date: Mon, 04 May 2009 21:29:33 -0400
- Subject: Re: Implementing BLOCK
- References: <49D49315.2040807@domob.eu> <20090402160426.GA56482@troutmask.apl.washington.edu> <49D4ECDC.9000806@domob.eu>
Daniel Kraft wrote:
This is of course true; to maybe shed some light on the issue, I was
thinking about working on ASSOCIATE (which is F2003, IIRC) also.
However, it seems to me that there are more subtle things to get right
and details to acknowledge for ASSOCIATE than there are for BLOCK, and I
did pick BLOCK as a first proposal because this seems to be easier and
less time-consuming to get working; unfortunatelly I've not much time at
the moment. Additionally, I'm currently holding a "for fun" Fortran
class at my university, and there already were some people asking
whether it was possible to declare variables with limited scope--because
of this and simply my (not very exhaustive) thoughts about the issue I
came to the idea that BLOCK might also be more interesting to users than
ASSOCIATE.
I've started using ASSOCIATE in my own code (Monte Carlo molecular
simulation), because ifort supports it. For obvious reasons, I don't
use BLOCK. I might or might not use it if it were available.