This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: DEFAULT_RECL too low?
- From: FranÃois-Xavier Coudert <Francois-Xavier dot Coudert at lcp dot u-psud dot fr>
- To: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>, gfortran <fortran at gcc dot gnu dot org>
- Date: Thu, 03 Mar 2005 10:19:06 +0100
- Subject: Re: DEFAULT_RECL too low?
- Organization: Laboratoire de Chimie Physique
- References: <42258BE7.7010907@lcp.u-psud.fr> <20050302180143.GA45397@troutmask.apl.washington.edu>
I think Steve forgot to Cc his answer to the gfortran mailing-list:
I don't know why one would choose 10000, because most hard drives
are partitioned into 512 byte sectors and most OSs write in groups
of sectors. It would seem to me that we would want a number like
16K, 32K, or 64K. Note, an increase in DEFAULT_RECL would only
paper over the bug because one can eventually find an N for
ARRAY(N) that exceeds DEFAULT_RECL.
Yes, I agree with that. A real fix would be to enable us to allocate a
buffer with size 2*DEFAULT_RECL when we see the output doesn't fit in
DEFAULT_RECL, and so on. But I'm not ready to write such code, unfortunately.
Janne's answer which makes sense to me. I will propose such a patch after
testing (probably tomorrow).
FX