2015-03-09 14:43 GMT+01:00 Thierry Goubier <thierry.goubier@gmail.com>:


2015-03-09 14:39 GMT+01:00 Nicolai Hess <nicolaihess@web.de>:
2015-03-09 14:18 GMT+01:00 Thierry Goubier <thierry.goubier@gmail.com>:
in short, implementorsOf:/referencesTo:/ is broken (does not search over the class side of things).

Moreover, depending on the kind of RBEnvironment one is looking at:

allClasses
�� �� returns only the classes
or
�� �� returns classes and metaclasses


I reported a couple of issues at fogbugz related to this (actually about smalllint validations rules, but they
��depend on this change on RBEnvironments)
And I raised some topics about this on the mailing list.

Ok, then we know what was the reason.
��

I was not involved in the latest changes on RBEnvironment and don't have any project that
depends on RBEnvironment.

So I can *do* the fixes I proposed, but I will not make the decisions.
Just waiting for feedback.

Which issues are we talking about? I can have a look.

Some rules aren't applied on the classside anymore
Some RBBrowserEnvironments don't work well with Metaclasses

closed but related:
critic rules running on all methods of a class extended by a package (closed)
CriticBrowser MNU Trait>>methodClass (closed)

closed, related and part of the problem:
Rewrite lint rules to check classes and methods
allClassesAndTraits significantly different from allClassesAndTraitsDo:


And mailing list "discussions"
http://forum.world.st/RBEnvironments-allClasses-Traits-Meta-tp4796370.html
http://forum.world.st/What-is-a-Behavior-are-Traits-behavior-too-tp4804691.html



��
��
��
Of course, implementorsOf: is broken because allClasses does not returns the metaclasses.

I like it when people "normalize" things :(

He asked for feedback on this. And as far as I know no one complained.

Fair enough.

Really need to have more test cases to cover things like that :)


RBEnvironments are so nice, small but really powerfull (refactoring/searching/validation).
I would like to get this right.


��

Thierry