Re: sdiff code error

[Paul Edwards]

> > But if sdiff has an option to produce a side-by-side
> > difference without interaction, then that bit of functionality
> > should be there
> 
> Sure.  It's called "diff -y".  You don't need sdiff at all.

Ok.

> Also, if all you're worried about is supporting POSIX, then you don't
> need sdiff.  sdiff isn't required by POSIX.

I'm not trying to support Posix, I'm just trying to port some
incredibly useful tools (diff, diff3 and patch) to a C89
environment.

> >> Incidentally, the code in question was introduced in diffutils 2.7.2
> >> (published 1998-09-14) so it appears you're the first person since
> >> 1998 (and the first person ever) to build diffutils on a host without
> >> sigprocmask.
> >
> > First person to report the problem, anyway.  :-)
> 
> Actually I suspect you're the first person, period.  It's been many,
> many years since people were seriously interested in pre-BSD
> platforms.

Well, what has happened is that someone has written an emulator
for IBM mainframes.  That means you can run IBM mainframe
operating systems on your PC.  And there was a period of time
when IBM provided the operating system for free.  Those operating
systems are now available.  That means you can do development
work for the mainframe on your PC.  Unfortunately you are
restricted to 16 meg of memory, but at least registers are 32-bit.

So I was able to port and debug GCC 3.2.3 on this old operating
system, and then have it working on a modern mainframe.  I'm
doing the same with some of the tools.  I've already ported sed
and bison.  I'm mainly interested in tools that are useful for
development, so that people can improve on GCC and PDPCLIB
(the public domain C runtime library which works on MVS and
others).  I'll be pretty happy when that environment is available.
I really love C.  It is the universal language, like English.

BFN.  Paul.