Related articles |
---|
Fortran 2003 ALLOCATABLE extensions manishagg22@gmail.com (2006-06-15) |
Re: Fortran 2003 ALLOCATABLE extensions jvorbrueggen@mediasec.de (=?ISO-8859-1?Q?Jan_Vorbr=FCggen?=) (2006-06-19) |
From: | =?ISO-8859-1?Q?Jan_Vorbr=FCggen?= <jvorbrueggen@mediasec.de> |
Newsgroups: | comp.compilers,comp.lang.fortran |
Date: | 19 Jun 2006 09:34:39 -0400 |
Organization: | MediaSec Technologies GmbH |
References: | 06-06-053 |
Keywords: | Fortran |
Posted-Date: | 19 Jun 2006 09:34:39 EDT |
> Can u tell me the effects on compiler code (scope of effect) due to
> the implementation of allocatable attribute for derived type? ,also
> the major problems relating to the implementation of the respective
> feature?
Well, practically all Fortran compilers on the market have had these
features before f03 was released, mostly because it was already laid
down in a TR several years before that. Thus, we know from experience
what can and will go wrong with compilers in this respect.
The problems mostly concern getting an allocatable's scope correct with
respect to allocation and deallocation. There are a lot of corner cases
you need to think about. The one that created most problems, I believe,
are dummy allocatable arguments with INTENT(OUT). Put that into several
levels of derived types, sprinkle with a POINTER here and there, and you
need to do some detailed and clear thinking with pencil and paper to get
things right.
Jan
Return to the
comp.compilers page.
Search the
comp.compilers archives again.