On Mon, Mar 14, 2011 at 7:02 PM, guilhermebla...@gmail.com < guilhermebla...@gmail.com> wrote:
> Hi all, > > I promise myself to not revamp this discussion again, but it wasn't me > this time! > > @Etienne: That RFC is outdated. > Since the last feedback form internals list, a lot of changes have > been made to that RFC. Maybe I should update it ASAP so you can > clearly understand what have changed to be compatible with current PHP > syntax. > > If you are interested, Pierrick moved all the recent developments to a > github repository, which can be reached here: > https://github.com/adoy/PHP-Annotations > > Take a look at some tests: > > https://github.com/adoy/PHP-Annotations/blob/master/tests/annotations/parser_021.phpt > > https://github.com/adoy/PHP-Annotations/blob/master/tests/annotations/ReflectionParameter_getAnnotations_003.phpt > > https://github.com/adoy/PHP-Annotations/blob/master/tests/annotations/ReflectionClass_getAnnotations_004.phpt > > Also, there's even an alternative patch that support positioned > parameters instead of named ones. > We just have to reach an agreement with what PHP core want. > > @Marcelo: While your proposal looks very good, it lacks of the support > to nested Annotation. > Consider how userland/framework would use your idea. For example, > Symfony2 supports validation of data inside classes inspired on > JSR-303 (Bean Validation). > > Symfony2 takes an advantage of a library Doctrine group (which I'm a > core member) created by parsing docblocks. When we created this > parser, I created this RFC with the good intention that PHP could > benefit of this known feature to enhance current userland > developments. > The first thing you need is your application still running ok with and > without comments. This already breaks all suggestions of creating a > PECL extension of docblock parser. > > > > I'd like to see what people think about it and make something "IN" on > next PHP major version. > > > now that the wiki is back and this was brought up in the 5.4 release planning, I think it would be a good idea to: - update the RFC to be in sync with the implementation - review the rfc and the patch itself - make the necessary modifications if necessary - decide whether we want this in 5.4 or not. Tyrael
