From: Matthew Kerwin <matthew@...>
Date: 2012-09-01T14:09:40+09:00
Subject: Re: Regular Expression Conundrums

On 1 September 2012 14:02, Stan Stanislaus <lists@ruby-forum.com> wrote:
> http://pastie.org/4643789
>
> Why, in Case 2, do I have to escape the asterisk twice to keep the
> program from thinking /\ is a substarter? Thank you all for your help! I
> really appreciate this.

Remember that the parameter you're passing in is a string.

In string literals, backslashes that precede any non-special
characters (such as asterisk) are essentially ignored, so "/\*/" is
the same as "/*/" (i.e. three characters: slash, star, slash).

  Regexp.new "/\*/" # => /\/*\//
  Regexp.new "/*/" # => /\/*\//

That resulting Regexp pattern is an escaped forward-slash, a star, and
an escaped forward slash; so it will match zero or more
forward-slashes followed by another forward-slash.  That's the same as
%r{/*/} or %r{/+}
Thus the numbers you get back are actually just the positions of every
sequence of slashes in the string.  It doesn't think "/\" is a
substarter, just "/"

Putting a second backslash in the string literal means you have four
characters in your string: slash, (escaped) backslash, star, slash;
which interpreted as a Regexp means: a slash followed by an (escaped)
asterisk, followed by a slash.

You could use Regexp#quote to take away the special meaning of the
star in your string; e.g.

[line 20]
  regex = Regexp.new( Regexp.quote regex_string )

and/or you could allow the regex_string parameter to be regexp object.
 That way you could call, for example:

  substarters = this_room_data.all_occurances "/*/" # no need to
escape regex characters
  substarters = this_room_data.all_occurances %r{/\*/} # no need to
escape string characters

No double-escaping; twice as much win.


-- 
  Matthew Kerwin, B.Sc (CompSci) (Hons)
  http://matthew.kerwin.net.au/
  ABN: 59-013-727-651

  "You'll never find a programming language that frees
  you from the burden of clarifying your ideas." - xkcd