| Steven G. Kargl 2007-11-26, 7:17 pm |
| In article <474b004d$0$8877$4c368faf@roadrunner.com>,
"Craig Dedo" <cdedo@wi.rr.com> writes:
>
> I agree with both Mike Prager and Gary Scott for the reasons that they
> mention. Since backslash is a printable character in the Fortran charactder set
> in Fortran 2003, any other interpretation of backslash violates section 1.5,
> "Conformance", of the Fortran standard. The following sentence from Fortran
> 2003 1.5, paragraph 4 [2:35-36], is particularly relevant:
>
Which is completely and utterly irrelevant for a Fortran
95 compiler where backslash is not a member of the Fortran
character set. There is an enormous difference between a
Fortran 95 compiler that offers a few Fortran 2003 features
and a Fortran 2003 compiler that is missing a few features.
> "A standard-conforming processor may allow additional forms and relationships
> provided that such additional forms and relationships do not conflict with the
> standard forms and relationships."
According to the above sentence, gfortran is a conforming Fortran 95
compiler because, well surprise, backslash is not a member of the
Fortran character set.
FWIW, the gfortran developers appear to be on the verge of
changing the default behavior for the 4.3.0 release. However,
this change will not suddenly give gfortran conformance to
Fortran 2003's Fortran character set.
--
Steve
http://troutmask.apl.washington.edu/~kargl/
|