From: "NARUSE, Yui" <naruse@...>
Date: 2010-11-08T18:04:34+09:00
Subject: [ruby-core:33103] Re: Suggestion for MatchData#first and #last

2010/11/8 Robert Klemme <shortcutter@googlemail.com>:
> On Mon, Nov 8, 2010 at 9:34 AM, NARUSE, Yui <naruse@airemix.jp> wrote:
>> 2010/11/8 Robert Klemme <shortcutter@googlemail.com>:
>>> Btw, I just notice this
>>>
>>> irb(main):016:0> /(?<f>fo+).*(\g<f>)/ =~ "foooobarfoo"
>>> => 0
>>> irb(main):017:0> $1
>>> => "foo"
>>> irb(main):018:0> $2
>>> => nil
>>> irb(main):019:0> f
>>> => "foo"
>>> irb(main):020:0> RUBY_VERSION
>>> => "1.9.1"
>>> irb(main):021:0> RUBY_PATCHLEVEL
>>> => 429
>>>
>>> I thought named group would have the same capturing semantics as
>>> unnamed groups but apparently the leading named group is not
>>> considered as capturing. �Is the behavior shown above intended?
>>
>> When you use named capture, you can't use normal parentheses.
>> It becomes (? .. ).
>
> I think this is not true, because as long as I do not use \g I can
> access the group via name *and* via index:
>
> irb(main):004:0> /(?<a>fo+)/ =~ "foobar"
> => 0
> irb(main):005:0> $1
> => "foo"
> irb(main):006:0> a
> => "foo"
>
> That's also in line with what the doc says
> http://www.geocities.jp/kosako3/oniguruma/doc/RE.txt
>
> Or did you mean "when using *calls* to named captures"? �Because
> that's where the difference comes in.

I mean, in RE.txt's words, named group.
When you use named group, (?<f>...), in a regexp,
(...) in the regexp is treated as (?...).

-- 
NARUSE, Yui
naruse@airemix.jp