From: Austin Ziegler Date: 2005-09-12T09:58:30+09:00 Subject: Re: new RCRs On 9/11/05, Eric Mahurin wrote: > I just wrote up several RCR's that I've been thinking about for > a while and have been discussed on ruby-talk: Quick Summary: > 318 easy way(s) to make a duck I'm strongly opposed to this one. David Black has nicely summarised something that closely resembles my opinion on this. The simplest way is "there is no duck." The next way is "try or try not." If you're wanting an easier way to make mock objects, that's something that can be done without modifying the language -- and should be done that way. But one can't "make a duck". Duck-typing is an *approach* to programming, not a typing system on its own. > 317 klass.from_s(str) or more generally klass.from_*(...) I'm also strongly opposed to this one. I think that the from_* is poorly considered and has been deconstructed sufficiently in its own thread. The more I use Ruby, the less I am convinced that a generic conversion mechanism will work or is even a good idea. If this *is* accepted and implemented (and I hope it isn't), I would want to ensure that missing conversion routines aren't necessarily the reported location of errors when it's not appropriate. > 316 arguments after a splat in a method call Matz has accepted this for Ruby 2. > 315 allow required arguments after optional arguments I like this idea, but I'd rather see it fixed up with fully named argument support rather than the meth(req-args, opt-args, *splat-args, **kwargs) approach considered now. > 314 concise expression debug/trace method IMO, this is something that make life much easier for me. -austin -- Austin Ziegler * halostatue@gmail.com * Alternate: austin@halostatue.ca