bschwart@elbereth.rutgers.edu (Barry Schwartz) (03/21/89)
I suppose I am going to get flamed for posting this. But here I am, and I must have my say or let my thoughts eat at me. Some of us could benefit greatly from a GNU project Fortran compiler. The compiler would have to be compatible with both the old and new standards. Is anybody working on such a compiler? If not, does anybody want to try? Don't look at me. The only "compiler" I ever wrote was a little FORTH system (as a term project), and also much of a modified fig-FORTH for TRS-80 Model I. But that is completely different from writing a FORTRAN compiler. --------- Please don't flame me too much, fellow C and C++ programmers. It's just that Fortran is the language of numerical analysis. The standard numerical packages are written in Fortran; the IEEE digital signal processing programs are written in old Fortran. Besides, however much I love C and C++, I also like Fortran. --------- A GNU Fortran compiler would be a downright neat thing, I bet. Barry Schwartz bschwart@elbereth.rutgers.edu
w-colinp@microsoft.UUCP (Colin Plumb) (03/23/89)
For people on the mailing list: I tried to cancel my message, but it got out anyway. The GCC talk is in Boston. The Speaker is from Stanford, but won't be speaking there. Lest anyone else be as confused as I was. I will not try to think at 4 in the morning. I will not try to think at 4 in the morning. I will not try to think at 4 in the morning. -- -Colin (uunet!microsoft!w-colinp)
johnl@ima.ima.isc.com (John R. Levine) (03/23/89)
In article <23@microsoft.UUCP> w-colinp@microsoft.uucp (Colin Plumb) writes: > All you need is lots of understanding of GCC's innards, Fortran >parser technology (DO10I=1.10 vs DO10I=1,10 and all that), and a lot >of spare time. I have donated an F77 subset parser that does a reasonable job of tokenizing. (It's the one I offered in comp.compilers a few months ago.) Extending it to the rest of F77 is pretty straightforward. Then all you have to do is the other 97% of the work to handle the symbol table and semantics, glue it into the GCC code generator, and write the I/O library. Spare time, indeed. -- John R. Levine, Segue Software, POB 349, Cambridge MA 02238, +1 617 492 3869 { bbn | spdcc | decvax | harvard | yale }!ima!johnl, Levine@YALE.something You're never too old to have a happy childhood.