home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
OS/2 Shareware BBS: 10 Tools
/
10-Tools.zip
/
f2csrc.zip
/
f2csrc
/
mbox
< prev
next >
Wrap
Internet Message Format
|
1993-07-30
|
10KB
From dmg Tue Jul 11 08:08:46 EDT 1989
f2c's formatted I/O (at least the little I've tried)
seems to work with Sun's libF77 and libI77 on sol;
some list-directed I/O works, but not character
strings -- Sun's f77 issues different calls for them.
From gre Fri Aug 4 21:09:07 EDT 1989
f2c.news (dmg) Fri Aug 4 17:24:39 1989
Now available for "friendly use", f2c converts Fortran 77
source code to C. By default, it tries for bit-level
compatibility with f77. Usage:
f2c [flags] file ...
File names must end in .f or .F; corresponding .c files are
produced in the current directory (unless no file arguments
are given, in which case f2c acts as a filter). `man f2c`
describes flags and other details. Bug reports to dmg.
From dmg Sat Aug 5 13:21:08 EDT 1989
I've tweaked things in netlib/f2c so one can pass options by
having the first line after "execute f2c" begin with "c$f2c ",
followed by options. Source for F2cargs is /n/pipe/usr/dmg/f2c/F2cargs.c .
(This also involved tweaking /n/crab/usr/ehg/bin/netlib.f2c .)
From coma!ehg Sat Aug 5 16:30:00 EDT 1989
>From research!bowell!dmg 6 Aug 89 10:24:36 GMT
(bowell.1828) comet : /usr/bin/f2c
New flag -r8 promotes REAL to DOUBLE PRECISION, COMPLEX to DOUBLE COMPLEX
(a la Berkeley f77).
From coma!ehg Thu Aug 10 16:30:00 EDT 1989
>From research!bowell!dmg 11 Aug 89 06:39:59 GMT
(bowell.1841) comet : /usr/bin/f2c
Quietly allow names up to 8 characters long (except under -ext).
Complain about, then ignore arguments to a main program.
Under -u, complain just once about each undeclared variable.
From arpa!hilbert.cray.com!tam Wed Aug 9 20:26:04 CDT 1989
Received: from hall.cray.com by uc.msc.umn.edu (5.59/1.14)
id AA13772; Thu, 10 Aug 89 17:57:59 CDT
Received: from tmacd.cray.com by hall.cray.com
id AA09337; 3.2/CRI-3.10; Wed, 9 Aug 89 20:26:10 CDT
Received: by tmacd.cray.com
id AA05465; 3.2/CRI-3.10; Wed, 9 Aug 89 20:26:04 CDT
Date: Wed, 9 Aug 89 20:26:04 CDT
From: tam@hilbert.cray.com (Tom MacDonald)
Message-Id: <8908100126.AA05465@tmacd.cray.com>
To: asg%3k@arpa.att.com, dmr@research.att.com, ehg@research.att.com,
nls@research.att.com, pjw@research.att.com, wmc@research.att.com
Subject: f2c help
Hi,
I promise not inundate all of you with so much Email in the future.
We tried the 'f2c' translator (or at least Bertice Haines did) and
she sent me the following Email.
======================================================================
Tom
It doesn't look like I'm going to get very far with the fortran
to C translator. It considers variables longer than six characters
to be errors and truncates them. It also mentions libraries that
we don't have and need a header file called f2c.h
Bertice
From: netlibd@arpa.att.com
Date: Tue, 8 Aug 89 22:02:21 EDT
To: arpa!bedlam.cray.com!bjh
Subject: filter f2c
Status: R
/* -- translated by f2c (version 0.6, all of Fortran 77).
You must link the resulting object file with the libraries:
-lF77 -lI77 -lm -lc (in that order)
*/
#include "f2c.h"
========================================================================
I was wondering if anyone else had complained about these issues.
Our Fortran compiler cft77 allows longer variable names. Also
can I get a copy of the "f2c.h" header. And finally, can anyone
tell me what the libraries libF77.a and libI77.a contain. I'd
really like to use 'f2c' as a way of finding out where the Standard C
compiler is deficient (performance wise). It also might help me understand
what I need to do to support your C environment.
Thanks for your help.
Tom MacDonald
tam@cray.com
uunet!cray!tam
From ehg Thu Aug 10 22:10:25 EDT 1989
From-: Eric Grosse ehg@research.att.com 201-582-5828
To-: tam@cray.com 3k!asg dmr nls pjw wmc dmg
Date-: Thu Aug 10 22:10:24 EDT 1989
Subject: Re: f2c help
(I'm sending this to the other recipients so they know they
don't have to reply. In general, the right fellow to contact
is David Gay, research!dmg. But he's going to be gone for a bit,
so I'll reply as best I can.)
From ehg Thu Aug 10 22:15:03 EDT 1989
From-: Eric Grosse ehg@research.att.com 201-582-5828
To-: tam@cray.com 3k!asg dmr nls pjw wmc dmg
Date-: Thu Aug 10 22:15:02 EDT 1989
Subject: Re: f2c help
Oops, the system truncated my message. Here goes another try...
[long names] I also consider Fortran names of more than 6 characters
to be errors. But David sent out mail this morning saying that f2c
would now allow names up to 8 characters; perhaps that will help.
But in the final analysis, this is Fortran77->C, not CFT->C.
[libraries and headers] About a week ago, David also updated the
index file in netlib to point to these supplementary files. Say
mail netlib@research.att.com
send f2c.h from f2c
send libf77 from f2c
send libi77 from f2c
and you'll get back the items you asked for. I know it is a little
startling to see the converted code calling the Fortran runtime
library. I understand that this policy was adopted in order that
the converted program would give output byte-for-byte identical with
the original Fortran. Of course, this is only true on classical
Unix systems running Feldman's compiler. Still, if you can match
these codes to your own runtime library closely enough, perhaps you
too can compare the Fortran and C output automatically.
Please pass along my best wishes to Bertice. She did an
excellent job for us while she was at Murray Hill.
Eric
From 3k!asg Fri Aug 11 16:11 EDT 1989
To: pipe!ehg
Subject: Re: f2c help
You can send your regards/complements directly to Bertice,
bjh@cray.com. Aaron
From arpa!bedlam.cray.com!bjh Tue Aug 15 09:15:20 CDT 1989
Received: from sc.msc.umn.edu by uc.msc.umn.edu (5.59/1.14)
id AA25045; Tue, 15 Aug 89 09:13:30 CDT
Received: from hall.cray.com (hall.cray.com.ARPA) by sc.msc.umn.edu; Tue, 15 Aug 89 09:53:40 cdt
Received: from pacific.cray.com by hall.cray.com
id AA18668; 3.2/CRI-3.11; Tue, 15 Aug 89 09:15:25 CDT
Received: by pacific.cray.com
id AA16716; 3.2/CRI-3.11; Tue, 15 Aug 89 09:15:20 CDT
Date: Tue, 15 Aug 89 09:15:20 CDT
From: bjh@bedlam.cray.com (Bertice Haines)
Message-Id: <8908151415.AA16716@pacific.cray.com>
To: dmg@research.att.com, ehg@research.att.com
I think I now have all of the pieces needed to convert some of our
cft77 regression tests using f2c. Thanks for your help.
Bertice
From coma!ehg Fri Sep 1 04:30:00 EDT 1989
>From research!bowell!pg 1 Sep 89 15:29:25 GMT
(bowell.1861) comet : lp
shipped /usr/spool/lp/devices /usr/spool/lp/process/generic /usr/spool/lp/spooler/generic /usr/spool/lp/spooler/dkgeneric
Now knows about "lino".
Some general cleanup.
>From research!bowell!dmg 1 Sep 89 15:45:53 GMT
(bowell.1862) comet : /usr/bin/f2c
Make A(min(i,j)) work when A is an array.
Add 7 and 8 character C keywords to the list of names
that have an underscore appended.
Accumulate block data in /tmp rather than the current directory.
Issue an error message for LOGICAL*1 (just one per LOGICAL*1 type
stmt, rather than one per variable in the stmt), and treat LOGICAL*1
as LOGICAL rather than declining to produce any C output.
Allow INTEGER variables to be used as formats (a la f77 and FORTRAN 66).
>From research!bowell!norman 1 Sep 89 17:35:07 GMT
(bowell.1863) comet : uucp/Systems.dk
september NAC shipment.
if you are in the habit of snitching Systems.gen as well,
pick up a new Dialcodes file; prefix DLl has been added.
>From research!bowell!doug 1 Sep 89 20:50:49 GMT
(bowell.1864) comet : man.macros
tweak to avoid hyphenating certain literals
------m dmg nls Sat Sep 16 00:19:45 EDT 1989
To-: dmg nls
The message
Subject: execute f2c
x=1
end
is, according to the RFC822 mail standard, just one (continued)
header line, and the message body is empty. If people complain
to you that they're getting back empty C, that's probably the cause.
From coma!ehg Tue Oct 3 04:30:00 EDT 1989
>From research!bowell!dmg 3 Oct 89 14:34:35 GMT
(bowell.1934) comet : /usr/bin/f2c
Fix bugs in character DATA for noncharacter variables
involved in EQUIVALENCE.
Portable initialization of noncharacter variables
with character data.
IMPLICIT NONE is now treated as IMPLICIT UNDEFINED(A-Z).
New option -!I disables includes (for netlib).
Bail out when given an invalid option.
From cs.utk.edu!dongarra Sat Aug 25 17:37:05 EDT 1990
Received: by pyxis; Sat Aug 25 17:36 EDT 1990
Received: by inet.att.com; Sat Aug 25 17:36 EDT 1990
Received: by dasher.cs.utk.edu (4.0/2.6c-UTK)
id AA02469; Sat, 25 Aug 90 17:37:05 EDT
Date: Sat, 25 Aug 90 17:37:05 EDT
From: dongarra@cs.utk.edu
Message-Id: <9008252137.AA02469@dasher.cs.utk.edu>
To: ehg@research.att.com
Subject: Re: netlib news
Eric,
I copies most things from /netlib/temp. I think I screwed up with /netlib/go,
so when you get a chance please put it back. I removed all files with
the exception of admin. I haven't copied that yet, since I need to check
the index and LIBS before coping.
Does the newest version of netlib automatically do the f2c function?
I don't think I have the latest.
Jack
------m dongarra Sun Aug 26 17:29:12 EDT 1990
I don't remember what was in temp/go. I'm running a complete check now
and that will cover any problems with go.
The f2c execution stuff here has lots of dmg (=David Gay) specific
file names, so when I copied the files to Oslo, I didn't bother to
install the executables. I get the impression that most people
these days want to run it on their own machine anyway. The main
motivation for sending a file to netlib would be as a quick test-run
or as a has-this-bug-been-fixed-yet test. For those purposes, it
is probably best that they send the files to David's latest version
anyway.
Best wishes
Eric