On Thu, Nov 13, 2003 a.d., Rafael Garcia-Suarez wrote: > Well, the original problem is at > http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/2003-10/msg01449.html > > In fact what was wanted (and implemented by this patch) is the ability to > access the PerlIO API early during the cloning -- to be used by debugging > flags (-D) for example. The whole idea of using PerlIO printing functions (which do malloc a lot themselves and mess with perl's internal state) to track leaking mallocs is at least curious :-) Perl is not (yet) an OS: one could use something external to clobber the console ex. warnx(3). IMHO, that change is badly breaking some bits of Perl functionality (using a CLONE method from an XS module) to unbreak something that's broken by definition. Regards, AdiThread Previous | Thread Next