On Thu, Oct 21, 1999 at 08:26:42AM +0200, Johan Vromans wrote: >On Thu, Oct 21, 1999 at 12:58:49AM -0500, Ed Peschko wrote: >> On Wed, Oct 20, 1999 at 08:59:08PM -0700, Chip Salzenberg wrote: >> >Have you considered >> > export PERL5OPT='-Mstrict -MDBI -MIO -MData::Dumper' >> >> yes, about five minutes ago... Now if you could compile a default value for this >> into a perl executable, I'd be in heaven... > >I think you'll regret it very fast. It's a long way down to earth... >Doing so will effectively create your personal (personalized) perl. Any time >you show someone a neat and handy perl script, the other person will not >be able to reproduce your results. After a while you forget about the >personalized setup, and you'll start wondering why other people's scripts >do not run as expected. Or why your script does not run on another system. >Or with a newer Perl version. No, I've thought about it quite a bit, and I'd definitely *not* regret it. Yes, there's a big 'user beware' element here, but when your scripts are 50% use, 50% 'real code' there comes a point in which it makes sense to throw away the whole visibility thing and simply do what is easiest. As long as it is optional that is. I hate to say it (I've said it before and I'll say it again), but rebol is considerably easier at certain things than perl for this very reason - adding such an option seems a good balance between the 'god damnit the core is sacrosanct' people and the 'why isn't this *easier*' folks. And if your install script is portable (configure wise) with a preconfigured PERL5OPT value, then you have no internal portability problems. Now, I admit that this poses a bit of a problem with CPAN - but I'd make it an (enforced) general rule that CPAN modules have to have PERL5OPT turned off. Considering that this functionality has existed for some time now without problems on CPAN, I don't think it would be such a big problem. EdThread Previous | Thread Next