[ruby-core:28026] Re: RCR: RUBY_VERSION_INT

Hugh Sasse:

> On Wed, 3 Feb 2010, Shot (Piotr Szotkowski) wrote:

>> I don=E2=80=99t like the =E2=80=98one digit=E2=80=99 policy, but this is=
 precisely why it was
>> instituted =E2=80=93 to have string comparisons of version numbers =E2=
=80=98just work=E2=80=99.

> Having a Version class would solve this

I agree =E2=80=93 I really don=E2=80=99t like artificial policies introduce=
d just for
side-stepping implementation details (but I also agree with Nobu that
checking for features is the =E2=80=98right way=E2=80=99 in most cases =E2=
=80=93 especially
when such features can be added dynamically, regardless of Ruby
version/implementation).

>  RUBY_VERSION > '1.9.0.9'
> ArgumentError: comparison of Version with String failed

If we were to have a Version class, there=E2=80=99d be no reason why it cou=
ldn=E2=80=99t
coerce Strings to Versions automagically on compare (it=E2=80=99d be the on=
ly
sane thing to do, IMHO, if RUBY_VERSION were to be a Version object).

=E2=80=94 Shot
--=20
The endless cycle of hope and disappointment has hardened cynicism.
I am afraid that we have reached the point where talking is useless.
                       [Jonathan Edwards, The Future of Programming]