Front page | perl.bootstrap |
Postings from July 2000
Dandy! I like this idea. As stated in an earlier message, I'm quite concerned about the possibility of ending up with a zillion versions of perl, and no clue as to which ones are 'real' perl. This proposal will at least weed out the 'cut down' versions, since they don't meet a minimum standard. However, it won't do much for proving that an extended version isn't 'real' perl. After all, the "trip tests" won't be able to look for non-standard extensions. At least not reliably, without deliberate inclusion of some method of listing all the extensions currently included within the Perl Core. I don't think I want to do that, since a skilled developer could hide it anyway. However, something like that might be useful for scripters, in conjunction with the 'require' command, to ensure that a person attempting to run the script knows what features it requires. (Turn it into a useful feature, and people won't be so cranky about being 'forced' to do things.) Andrew Greene wrote: > > Perhaps we could adopt a TeX-like strategy: > > * Any executable called "perl" must be 100% conformant, standard, > passes the "trip tests." > > * Any modified version of perl -- whether because certain features > have been omitted or because certain modules are "compiled in" -- > must be called something else. "siteperl" or "microperl" or > what-have-you.