Re: systems requiring exit?

[Howard Chu]

Matthew Woehlke wrote:
> Howard Chu wrote:
> > Paul Eggert wrote:
> > > "Ilya N. Golubev" <address@hidden> writes:
> > > > Have no systems with broken `return' at hand,
> > >
> > > That's the fundamental problem.  These systems are _ancient_ -- they
> > > all predate C89 -- and they are so old that nobody uses them any more.
> >
> > As I recall when porting a number of packages to OS/390 and z/OS, all 
> > of configure's conftests failed using "return" (SEGV) and I had  to 
> > change them all to "exit"...
>
> Hmm, mainframes? I want to say VMS also uses 'return 1' for success 
> (or maybe it even needs exit; I'm not the one that has to deal with 
> our VMS systems). But as I recall VMS doesn't have a POSIX shell 
> either. Is there any VMS support that anyone is worried about?
Yes, there is a POSIX environment available on the mainframes. It's not 
the most stable thing in the world (return from main() shouldn't cause a 
SEGV, obviously) but it's possible to do useful work in it. I ported all 
of BerkeleyDB, OpenSSL, Cyrus SASL, Heimdal Kerberos, and OpenLDAP, and 
it all ran with reasonable performance.

--
 -- Howard Chu
 Chief Architect, Symas Corp.  http://www.symas.com
 Director, Highland Sun        http://highlandsun.com/hyc
 OpenLDAP Core Team            http://www.openldap.org/project/