Front page | perl.perl5.porters |
Postings from July 2009
* Paul Fenwick <pjf@perltraining.com.au> [2009-07-05 15:05]: > So regardless of whether or not %^H triggering object > destruction was an accidental or intentional feature, it's > now being depended upon by a lot of code, including > Catalyst::Runtime. Your defensiveness is not becoming. Just because it got used by a bunch of code does not make it a good idea, any more than `my $foo if 0` was, which has certainly seen far wider use. * Paul Fenwick <pjf@perltraining.com.au> [2009-07-05 15:50]: > I certainly didn't expect it to move to "let's boot autodie > from the core" territory. That rather surprised me too. > I don't want to see autodie get pulled from the core +1 > I'm happy to document the bug and its workarounds I think at this this time this is the best option. Next on the agenda: an actual discussion about the needs of autodie and similar modules. The current reliance on unspecified behaviours, namely incidental stringification of values at scope boundaries and respect for magic attched to a special global, which strike me as very shaky, seems indicative of the fact that %^H as provided is underpowered for practical use. Let’s figure out what needs have to be covered and how an explicit, sane and supported API might look like. It is reliance on implementation accidents that makes the DarkPAN (in fact, the mere DarkPAN *meme*) a problem of the magnitude it currently represents. Regards, -- Aristotle Pagaltzis // <http://plasmasturm.org/>Thread Previous | Thread Next