Re: BC breaks: Cache_Lite, PHP_Documentor

> > I'm a bit reluctant to leave as is for too long (read: years) - both >> packages are fairly popular, and if PHP 5.4/6 kills off the methods; the >> migration path for a heck of a lot of other users is much more painful >> (Cache_Lite -> Zend_Cache, PHPDocumentor -> ... doxygen?). >> > > You are missing my point again: the migration path for users is to stop > relying on magic_quotes_runtime and switch it off. The migration path for packages is to quit calling > set_magic_quotes_runtime() unconditionally (I now see that phpDocumentor > does that): this will prevent errors for people who are not relying on > deprecated functionality. > > Agree, but... how do we get the magic quotes people to fix their environments without raising errors (and thus introducing a BC break)? I'm talking the ancient, PHP4, there's no E_DEPRECATED raised folks as opposed to the PHP 5.* folks. All one of them. At what point do we go "oh nuts to those PHP4 users on http://stopphp5.org/, they keep posting about python anyway, let's make a minor BC break to make it visible"? The don't-leave-it-years desire - yes, we could conditionally check things; but I imagine in the next major version; get_magic_quotes functionality would probably go the way of the dodo too - so we end up having to do this dance again for the PHP 5.* users.