[Info-vax] Looking into C-include files on VMS

[Bob Koehler]

In article <4af3051e$1 at news.langstoeger.at>, peter at langstoeger.at (Peter 'EPLAN' LANGSTOeGER) writes:
> In article <7lg60tF3e2em8U1 at mid.dfncis.de>, js at cs.tu-berlin.de (Joerg Schilling) writes:
>>I don't know what has been tested in 1992. Today, it is impossible for VMS
>>to get UNIX branding as long as system() does not call a POSIX shell but
>>the VMS command line interpreter. open(2) may be another problem.
> 
> Are you sure about that?
> Is system() always calling DCL or does it perhaps depend on the user's
> UAF setting (CLI: DCL) and so this road block would be non existing
> (if correctly configured - btw what would this be CLI: PSH or CLI: BASH)?

   I think system() will drop into whatever CLI the parent process has
   mapped.  Which means that detached processes without a CLI mapped
   won't be able to use system(), just like they cannot use lib$spawn.

   If I ever get back to playing around with GNV again, I'd like to
   look into porting pdksh; which is an imperfect implementation of
   ksh, but should be a good starting port for a POSIX shell.

   But does UNIX branding require only a POSIX shell, or does it actually
   require certain shell utilities, like ls?  Since Linux depends on
   gnu for these, there should be code in either GNV or gnu repositories
   on which this could be based.