Either this is a bug, or I’m missing something:
irb(main):001:0> /^[aeiou]/.match(“foo\noooh”)
=> #MatchData:0x81d588c
irb(main):002:0> /^[aeiou]/m.match(“foo\noooh”)
=> #MatchData:0x81d2cf4
Shouldn’t the second one return nil?
ruby 1.8.1 (2003-12-03) [i386-netbsdelf]
Thanks.
–
Ceri Storey cez@necrofish.org.uk
Don’t mind me, I’m an idiot–I wanted ‘\A’ instead of ‘^’.
(Thanks Jason)
On Wed, Dec 10, 2003 at 01:56:43AM +0900, Ceri Storey wrote:
Shouldn’t the second one return nil?
–
Ceri Storey cez@necrofish.org.uk
I believe that Ruby’s version of /m just makes . match \n (which it normally
doesn’t); the ^ and $ always match around \n. If you want to match the
beginning or end of the entire string use \A instead of ^ and \Z instead of
$.
-Mark
On Wed, Dec 10, 2003 at 01:56:43AM +0900, Ceri Storey wrote:
Either this is a bug, or I’m missing something:
irb(main):001:0> /[1]/.match(“foo\noooh”)
=> #MatchData:0x81d588c
irb(main):002:0> /[2]/m.match(“foo\noooh”)
=> #MatchData:0x81d2cf4Shouldn’t the second one return nil?
You’re not an idiot. 
The /m modifier in Perl does do exactly
the inverse of what you thought it did in Ruby - make ^ and $
match around \n’s (in Perl they default to only the beginning and
end of the string). The various flavors of regular expression
floating around can be confusing, and a lot of Rubyists
start out in Perl, which doesn’t help in this instance.
The Perl->Ruby mapping looks like this:
Perl = Ruby Meaning
(no modifiers) (not supported) ^/$ at ends of str
/m (no modifiers) ^/$ around \n
/ms /m . matches \n
Then there’s Ruby’s /s, which doesn’t affect . or ^/$, but adds support
for matching multibyte characters (which you might expect to be /m, but
I think /s stands for SJIS).
-Mark
On Wed, Dec 10, 2003 at 02:03:41AM +0900, Ceri Storey wrote:
On Wed, Dec 10, 2003 at 01:56:43AM +0900, Ceri Storey wrote:
Shouldn’t the second one return nil?
Don’t mind me, I’m an idiot–I wanted ‘\A’ instead of ‘^’.