Re: compat_bsdos

[steve%prd.co.uk@localhost (Steve Blinkhorn)]

This is for some retrospective analysis for academic publication.  It's
all software I acquired long ago, compiled for BSD/OS without problem
after small tweaks to I/O channels.  It's originally Fortran IV, and
comews from a number of highly reputable (in termws of code
portability) sources, such as Bell Labs.  There seems to be no -f66
option for f2c or fort77.  In the process of failed compilation I see
some common issues: inconsistent calling sequences, where a subroutine
parametr has been translated to a real in one place and an integer in
another, recommendations to use non-existent options as a remedy,
complaints bout typing in general.

The binary compatibilties were the thing that really attracted me to
NetBSD when BSDi folded.  But they've let me down before - the Linux
compatibility have lagged behind what colleagues in other institutions were
working with, so I have from time to time to run a Linux box.  

--
Steve Blinkhorn <steve%prd.co.uk@localhost>
You wrote:
> 
> 
> > On Jul 21, 2020, at 6:19 AM, Steve Blinkhorn <steve%prd.co.uk@localhost> wrote:
> >=20
> > I have suddenly discovered the absence of compat_bsdos, having just
> > completed the replacement of all our old x86 machines with amd64s.  Is
> > it irretrievably lost?  I have a mass of software compiled originally
> > for BSD/OS, and quail at the magnitude of the task of recompilation,
> > particularly since much of it has Fortran source, and f2c-f77 has got
> > fussier about common Fortran practices of old.  The thought of fixing
> > the kinds of issues that may arise (I've tried one program, and it
> > hasn't been fun - it's still not working) does not fill me with
> > anticipatory pleasure.
> 
> I can't imagine that it would be too much work.  It should be pretty =
> similar to 4.3/4.4, and we have compat layers for those vintages on =
> various platforms.
> 
> -- thorpej
> 
>