Re: deep investigation of PHP_Callback

Hi, Lukas Kahwe Smith wrote:

What Bertrand and Alexey were worried about was that this package would become the defacto standard for all callbacks implemented in PEAR, which would quickly add up to quite a lot of overhead. So the key things for me is to ensure that people realize when it makes sense to use this package and when it doesn't. This also ties in to the general informal guideline in PEAR that you do not put dependencies in order to save a minimal amount of code (let's but the number somewhere around 50 lines). This general guideline would probably prevent the worst of abuses of PHP_Callback.

I won't speak for Bertrand since we didn't discuss this with him, but my main reasons against was the following: 1) We accept packages that solve specific user problems. I don't see a problem being solved here in the first place. 2) Traditionally PEAR didn't accept packages that just wrap native PHP functions in fake OO code. If that was not the case we'd already have PHP_Callback / PHP_String / PHP_Sprintf and whatever else. 3) We all fondly remember PEAR_Error. 'nuff said. Also this whole discussion wasted much more of everybody's time than writing this gem of a package did. I won't repeat this error and will in similar cases wait with arguments until the voting phase. That'll be much more fun.