At 07:49 PM 4/15/1998 GMT, Phillip Helbig wrote:
>From: SMTP%"[log in to unmask]" 15-APR-1998 19:47:36.29
>
>> Once one has the subset definition, it will be natural to have flag
>> settings that either ignore the subset, warn of subset violations, or
>> reject subset violations. Many F90 compilers already have such flags
>> with regard to non-standard features, such as REAL*8, and perhaps
>> even deprecated features. such as ending a DO loop with something
>> other than END DO or CONTINUE.
>
>I believe standard-conforming compilers are REQUIRED to have such flags.
>
>> I.e., an ELF or F syntax checker. The advantage to going this route
>> is that, in principle, each compiler doesn't have to implement it
>> itself.
>
>Yes, and it could be written completely portably in a high-level
>language (like a Fortran95 subset). The FORCHEK syntax checker is, I'm
>pretty sure, written in Fortran90. ftnchek is written in C.
>
snip
I think I remember David Epstein telling me that the F compiler
was written in F, hence I infer that it should be compilable on any
platform with an F90 compiler.
How about a comment from Imagine1 on this? If that is
the case, it should be simple to make available at least the front-end
syntax checker of F for any F90 platform.
ralph frisbie
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|