negative look behind with unknown number of spaces

Question

Using c# regex I'm trying to match things in quotes which aren't also in brackets while also ignoring any white space:

"blah" - match
("blah") - no match
( "blah") - no match
(  "blah") - no match

I've got (unescaped):

"(?<=[^(]s")(.*?)"

which works with the first three but I can't work out how to deal with more than one space between the first bracket and the quote. Using a + after the s is the same result, using a * means both the last two match. Any ideas?

Answer 1

Look behinds need a fixed width, but you might be able to get there with the expression below. This assumes no nesting.

/G                 # from the spot of the last match
  (?:               # GROUP OF: 
     [^("]*           # anything but open-paren and double quote.
     [(]              # an open-paren
     [^)]*            # anything but closing-paren
     [)]              # a closing-paren
  )*                # any number of times 
  [^"]*             # anything but double quote

  "([^"]*)"         # quote, sequence of anything except quote, then ending quote
/x

Answer 2

In PCRE as I know it, lookbehinds have to be fixed-width. If that remains true in C#'s PCRE engine, then you're not going to be able to do it the way you're trying to.

Answer 3

This should work:

/(?<![^(s])s*"([^"]*)"s*(?![s)])/

• The first (?<![^(s]) asserts that there is no whitespace or left parenthesis before the string.

• Then s* will match any number of whitespace characters.

• ("[^"]*") will match a quoted string, and capture it's content.

• s* will match any number of whitespace characters.

• Last, (?![s)]) will assert that there is no whitespace or right-parenthesis following.

Together they make sure that all the whitespace is matched by each s*, and that they are not bordering a parenthesis.