[ruby-talk:00777] Re: Ugly behavior using /<r1>/ .. /<r2>/ cons truct.

On 31 Aug, Yukihiro Matsumoto wrote:
> Hi,
> 
> In message "[ruby-talk:00768] Ugly behavior using /<r1>/ .. /<r2>/ construct."
>     on 99/08/30, [e-mail] <[e-mail]> writes:
> 

[...]

Thanks for your explanation! Now I have understood the difference :-)
From that point of view, the current behavior should be correct! I could
use the `...' operator. Then it works like I want.

BTW: It is more correct than under Perl 5.005_02 ;-))))))))))))

Although my proposal was issued from a misunderstanding, I think it
would nevertheless, be nice to have it implemented ;-)

> |My proposal would be, that in such `..' construct, the variables `$n'
> |should only be set to the captured subpattern of the *matched* regexp.
> |The one that *doesn't* match, should not reset them! Only if both
> |doesn't match, they should be reset to `nil'.
> |
> |What do you think?
> 
> Let me take time to think about it.  Suggestion, comments are welcome.

Of course! :-)))

> 
>                                                         matz.

\cle