From: Robert Feldt <feldt@...>
Date: 2001-06-18T07:37:09+09:00
Subject: [ruby-talk:16580] Re: Trapping method defs

On Mon, 18 Jun 2001, Robert Feldt wrote:

> but you'll have to reinstall the singleton_method_added in
> singleton_method_added if someone tries to redef
> Object.singleton_method_added. And then they can probably use undef_method
>
But it won't work since we wont be notified if singleton_method_added is
redefined. At least I can't find a way... Anyone feeling the challenge?

Matz: What can and cannot be trapped seems somewhat arbitrary. Any plans
on redesigning the trapping/callback/trace_func related methods to give a
more detailed and unified access to events? One thing I'd like to see
would be sticky callbacks that must be explicitly taken away; if not they
can't be overridden by other uses etc. But there are probably some good
reasons for not doing such a thing (backwards compatibility for
one...). But maybe for Rite?

/Robert