home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.lang.fortran:3565 comp.unix.ultrix:6919
- Newsgroups: comp.lang.fortran,comp.unix.ultrix
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!usc!sdd.hp.com!ux1.cso.uiuc.edu!news.iastate.edu!cfrandal
- From: cfrandal@iastate.edu (Charles F Randall)
- Subject: Re: fsplit problem with Ultrix 4.2c and Fortran 3.2
- Message-ID: <1992Sep15.171027.23584@news.iastate.edu>
- Keywords: fsplit
- Sender: news@news.iastate.edu (USENET News System)
- Organization: Iowa State University, Ames, IA
- References: <15452@umd5.umd.edu>
- Date: Tue, 15 Sep 1992 17:10:27 GMT
- Lines: 28
-
- In article <15452@umd5.umd.edu> kathy@umd5.umd.edu (Kathryn M. Orhelein) writes:
- >When it is fsplit (into 20 routines), each compiled with f77 -c program.name,
- >and put together with f77 -o test2 main000.o ....
- >it gives a different set of results (consistently).
- >The system is a DS 5000/200 running Ultrix 4.2c and DEC Fortran V3.2.
-
- Look at all calls to subroutines and functions and examine the type
- and size of the arguments that are passed and those that are
- expected. I'll bet you'll find some pretty gross mismatches.
-
- The DEC Fortran compilers are able to try to "guess" at the subroutine
- interfaces when all units are compiled together. They can't "guess"
- when compiled separately. (Any DEC people out there care to comment on
- how this is done?)
-
- Compile the source with the "-V" option to get the compiler listing
- file. There'll be information there that tells you what the compiler
- thinks the size and type of all variables are. You should be able to
- use "ftnchek" to help you out also.
-
- FYI,
- Randy
-
- Charles F. Randall IV | _ | "Are you kidding? I study
- Aerospace Engineering | == \---------.., | from crisis to crisis!"
- Iowa State University | ==___-----_____> | -anonymous student
- cfrandal@iastate.edu | _____b________d___ | <== ascii space shuttle
-
-