Re: Mail/mimeDecode.php

[logenews1@morelogs.de (Marc Logemann)] > Hi, > > >I don't like userland implementations if there are > >already existing extensions, but I have one suggestion: > >the corresponding userland implementation *MUST* give > >advice that there exists already an extension which does > >the same thing (and perhaps better) and that user should > >use the extension if it's possible for him. If not, he > >should use the userland implementation instead. > > This is what Richard wrote some minutes ago, by having > the same interfaces in the userland impl. the user > get "fast and native" parsing if he has the extension, > otherwise he will get a "not fast but also ok" > userland implementation. You could test this > during runtime and the user has no problems... I'm worried that having userland default implementations for PHP extensions may end up hurting the reliability we're trying to establish for PEAR. But let's try to get all the pros and cons on the table: +1 : people using "nazi ISPs" are given access to functionality they otherwise would not have there -1 : there's plenty of stuff in C extensions that can't be implemented in PHP -1 : userland implementations may be dreadfully slow -1 : userland implementations will most probably have different bug and feature sets than their C counterparts +1 : developing (and fixing bugs) in PHP is faster than in C I think the "different bug/feature sets" problem is the most serious one. - Stig -- Stig Sæther Bakken <ssb@alltheweb.com> Fast Search & Transfer ASA, Trondheim, Norway