The recent thread on autoloading charnames tangentially reminded me of a problem we have at $work. The issue is we want to pre-fork load as much code as possible, so that we don't have multiple children post-fork loading the same stuff. For us this is both a time optimization and a space optimization. (Because we preload when the software is not under load, whereas autoloading happens during a response.) I understand that autoloading code is powerful tool that has a lot of positive attributes. But it would be nice if there was an easy way to tell Perl and Module code to preload everything and forget lazy loading. An example of a challenging preloading scenario (doable, but a little "tricky") all the regex utf8 reference data. So if and when we add things to Perl that autoload we should always keep the need to preload in mind, and provide a way to do so. While it is off topic, certain modules, such as DateTime, are a real PITA in this regard. Just a thought, yves -- perl -Mre=debug -e "/just|another|perl|hacker/"Thread Next