LPM: perldoc perlerrors

Frank Price fprice at mis.net
Tue Aug 29 16:58:52 CDT 2000


On Tue, Aug 29, 2000 at 05:51:48PM -0400, Matt Cashner wrote:
> On Tue, 29 Aug 2000 ken.rietz at asbury.edu wrote:
> 
> > How about a compromise? Create something like a header file that defines
> > $SYSTEM_CALL_SUCCESS = 0;
> > $SYSTEM_CALL_FAILURE = 1;
> > for now, and use only the symbolic values in the code. Then change the
> > header for Perl 6.
> 
> we are talking about documentation here arent we? i wasnt talking about
> writing code dependant on that feature but rather documenting this
> feature.  because someone someday will be given a crap load of perl5 code
> that depends on system's return value and (if 6 changes this
> drastically) wonder what in the world their predessor (sp??) was
> thinking. (I speak from experience here but in relation to perl4-ish
> code.)  so documentation of this feature might be nice.  

Yes, I think we're just talking about documenting this feature, and as
has been noted, a good "NB: this feature will reverse itself in perl6"
is a good idea :-)

> but yeah, i
> agree. dont write production code that is dependant on a feature you know
> will probably change in the next few years.

Hmmm.  What if you have to write that code now?  Since perl5.6
currently returns 0 on success from a system() call, and given that
checking that return is a good idea ... what other option do I have?

I don't really want to get in a shouting match about this, just my
thoughts.  

-Frank.
-- 
Frank Price | fprice at mis.net | www.sxse.org/fprice/
GPG key: www.sxse.org/fprice/gpg.asc | E Pluribus Unix




More information about the Lexington-pm mailing list