From: Yukihiro Matsumoto <matz@...>
Date: 2011-09-05T15:48:19+09:00
Subject: [ruby-core:39273] Re: RubySpec vs CRuby's test/...

Hi,

In message "Re: [ruby-core:39271] Re: RubySpec vs CRuby's test/..."
    on Mon, 5 Sep 2011 13:48:46 +0900, Marc-Andre Lafortune <ruby-core-mailing-list@marc-andre.ca> writes:

|> * Core CRuby developers have used test/* for long time, so they are
|> ��more familiar to test/*
|>
|> Developers are often lazy and ignorant. ��Simple documentation for
|> RubySpec contributors may work. ��If it already exists, advertising on
|> it (or referencing it from w.r.o) may work.
|
|Indeed, I should have pointed out the historic aspec of test/*. It
|looks like you feel that, today and for the developers that can, many
|tests would ideally be added to RubySpec, is this what you mean?

Yes, but I don't want to make rubyspec full of implementation
dependent corner case messes, like test/*.  Maybe the strategy should
be extracting implementation independent specification from messy
test/*.

|Also, I'm sorry, I'm not sure what you mean by "w.r.o."? I'll not that
|the {Developer|Committer}HowTo wikis gives information on how to
|update and run the RubySpec and that http://rubyspec.org has extensive
|documentation.

w.r.o is www.ruby-lang.org, sorry for using uncommon abbreviation.

Updating test/* is very easy for core developers to update, comparing
to making a patch to rubyspec and sending pull request.  Considering
extracting strategy above, this situation may not be bad.

							matz.