Re: [RFC] Deprecate PEAR/PECL & Replace with composer/pickle

Hi! > I think there is value in reducing the number of ‘bundled’ things, > particularly given that Debian (and thus Ubuntu), Centos and OS X all > distribute PEAR/PECL as a separate package. On the Linux distro's > it’s a separate (but possibly “recommended” by php-cli etc) package. > On OSX there is /usr/lib/install-pear-nozlib.phar. Don't see how it matters what distros do. They certainly build with non-default options and can re-combine stuff as they want. If you use a distro it doesn't matter to you how PHP build behaves since you never interact with it. > However, I find the concept of then adding in Composer/Pickle to be a > very odd choice, and a poor one “politically”. What you mean here? Why is it odd, given the popularity of composer, which is I think an established fact? Of course, not everybody uses it everywhere - as each other tool, it is useful for soem specific scenarios. And what you mean by "politically"? > As someone else said, bundling composer sounds like an official > endorsement of PHP-FIG, which I suspect a number of people would have > issues with. No it doesn't. What it has to do with PHP-FIG at all? You can use composer and create composer packages without ever hearing of PHP-FIG, ever interacting with it or ever needing to. In fact, I don't think FIG is ever mentioned on the composer site, except when referencing documents like PSR-0. I almost no idea what "issues" with PHP-FIG are, but whatever they are, they are completely irrelevant to the question of whether composer is a useful tool. IMO, it is, but if people think it is not, we could have a vote and see. > On a purely technical level - it makes little sense. As demonstrated > above, a decent number of environments are not going to ship it as an > included tool anyway. Well, yes, of course - I thought that the whole question is that we should *start* providing it as an option when building PHP, of course we would not cover environments which don't do it. -- Stas Malyshev smalyshev@gmail.com