mike@gangue.uucp (Mike Norred) (02/10/85)
I am the proud maintainer of numerous fortran programs of varying vintage. Other than advice to convert them all to C, I would appreciate some information: While I am quite aware that "portable fortran" is an oxymoron of sorts, has anyone done work in this area. Something like an f77 version of "lint" would be a great help, even if it spoke just one dialect. Thanks in advance for any thoughts or suggestions you may have on this subject. Its probably best if you could reply directly to me, and I'll summarize to net.lang.f77. Thanks, Mike -- gangue!mike (Mike Norred) at MINEsoft, Ltd. ...ucbvax!nbires!gangue!mike
ndiamond@watdaisy.UUCP (Norman Diamond) (02/12/85)
> While I am quite aware that "portable fortran" is an oxymoron of > sorts, has anyone done work in this area. Something like an > f77 version of "lint" would be a great help, even if it spoke > just one dialect. For f66, there was a program called "pfort", and articles were published about the pfort verifier. I believe it pre-dated lint. It helped make "portable fortan" a fact rather than an oxymoron. It spoke the ANSI "dialect", and also did extensive checking such as for subprogram calls. I regret not knowing if there is a pfort f77 verifier. It would make someone eat his words. -- Norman Diamond UUCP: {decvax|utzoo|ihnp4|allegra|clyde}!watmath!watdaisy!ndiamond CSNET: ndiamond%watdaisy@waterloo.csnet ARPA: ndiamond%watdaisy%waterloo.csnet@csnet-relay.arpa "Opinions are those of the keyboard, and do not reflect on me or higher-ups."
dgary@ecsvax.UUCP (D Gary Grady) (02/14/85)
<>
The PFORT ("portable FORTRAN") verifier accepted a very restrictive
subject of FORTRAN (END= not allowed, Hollerith constants required
instead of quoted strings, etc.) and did a good job of checking type
compatibility in subroutine calls and COMMONs.
The most impressive thing about PFORT was the fact that it was written
in PFORT verified FORTRAN (!!!!!). Assembler routines were used to
detect end of file, since that was not a standard FORTRAN feature in the
66 language.
I would really like to hear of a similar portable verifier for portable
FORTRAN 77. In fact, that would be a nice thing to have for a lot of
languages.
--
D Gary Grady
Duke U Comp Center, Durham, NC 27706
(919) 684-3695
USENET: {seismo,decvax,ihnp4,akgua,etc.}!mcnc!ecsvax!dgaryminitab@uwstat.UUCP (02/16/85)
> > While I am quite aware that "portable fortran" is an oxymoron of > > sorts, has anyone done work in this area. Something like an > > f77 version of "lint" would be a great help, even if it spoke > > just one dialect. > > For f66, there was a program called "pfort", and articles were published > about the pfort verifier. I believe it pre-dated lint. It helped make > "portable fortan" a fact rather than an oxymoron. It spoke the ANSI > "dialect", and also did extensive checking such as for subprogram calls. > > I regret not knowing if there is a pfort f77 verifier. It would make > someone eat his words. The PFORT Verifier is table-driven, and is, itself, written in PFORT. When I got it (late 1970s) it cost $35 from Bell Labs, was unsupported (of course!) and could be reproduced (but not for profit) providing the unmodified code was provided and the original form of the copyright notice was attached. I was able to easily modify PFORT to accept statements allowed by a preprocessor we had (like ASSERT (logical expression)), and many F77 constructs were already supported (block IF, CHARACTER). I never saw a better checker of interprogram communication before or since (though I here the Perkin-Elmer "global optimizing" compiler is better). I think every Fortran shop should have PFORT and the related FORTLEX Fortran lexical analyzer. Many useful tools can be built from these two. Stephen L. Arnold, Joiner Associates Inc. Mail P.O. Box 5445, Madison WI 53705-0445 Phone 608 238-8134 Telex 650 110-6813 BITNET arnold@wiscpslb ARPA arnold%wiscpslb.BITNET@wiscvm.ARPA UUCP {allegra|cornell|purdue|ihnp4|burdvax}!psuvax1!arnold%wiscpslb.BITNET
jsdy@hadron.UUCP (Joseph S. D. Yao) (02/25/85)
> While I am quite aware that "portable fortran" is an oxymoron of > sorts, has anyone done work in this area. Something like an > f77 version of "lint" would be a great help, even if it spoke > just one dialect. While most compilers will compile a majority of standard code, especially if it has passed 'pfort', an ex-professor of mine did a study of Fortran implementations in the late 60's-early 70's time frame. He found that of some reasonable number of compilers sampled, there were n o s t a t e m e n t s that were the same in all of them! Only one statement came close -- minor mods to two compilers would make the CONTINUE statement the same across the board. And people complain about minor differences in C and UNIX! ;-) Joe Yao hadron!jsdy@seismo.{ARPA,UUCP}