Date: Wed, 14 Dec 1994 22:52:52 +0900
From: raraki AT human DOT waseda DOT ac DOT jp (Ryuichiro Araki)
To: djgpp AT sun DOT soe DOT clarkson DOT edu, gunter AT statsun DOT stat DOT ColoState DOT EDU
Subject: Re:  f2c
Cc: gunter AT yuma DOT ACNS DOT ColoState DOT EDU

>I'm using the 1992 version of f2c.  Is this the newest version? 

No. The newest version of f2c as of now is 941210.  You can find the sources 
at ftp://netlib.att.com/pub/f2c/.  You may need some patches to successfully
compile them with djgpp.  Details on request.

>How well does f2c work?

It depends on what you'll do with f2c. As far as I know, f2c works
satisfactory if your Fortran code doesn't use extraordinary syntax 
(including vendor-specific extentions).  I've ever converted/compiled 
several huge scientific programs and subroutine packages written in 
Fortran (e.g., SPIDER (Subroutine Package for Image Data Enhancement and 
Recognition, AIST-MITI, JAPAN), MOPAC, SPICE, etc.) to C using f2c/djgpp 
without any serious problems (well, there was a little bit of troubles, 
though :-)).

>Are there known problems or bugs?"  

Read the latest documents otained at netlib.att.com.

>How does f2c handle C array's?  Do I need to transpose them or 
>anything before I pass them as arguments?

Are you talking about the difference in the order of column and row 
in 2-dimensional arrays between C and Fortran? If so, don't worry. f2c
treat it adequately.

   ---
   Ryuichiro Araki
   raraki AT human DOT waseda DOT ac DOT jp