please explain this regular expression

Question

  ^[[:space:]]*@

I can't figure out what the [[:space:]]* means in the above regular expression. Please help, thanks!

It is the word. I saw this regex from other people's code, but cannot understand it. — Runner, May 22 '13 at 01:20
@mclaassen That is not related. This is a POSIX character class, supported by Perl in order to allow regex writers to better conform to Unicode. — Platinum Azure, May 22 '13 at 01:24
Take a look at http://stackoverflow.com/a/1968223/1890567 its just a character class — FDinoff, May 22 '13 at 01:25

score 9 · Answer 1 · edited May 22 '13 at 03:00

[:space:] is a POSIX character class which matches All whitespace characters, including line breaks in the word.

In other words [:space:] is identical to \s (since Perl 5.18^[1])

http://www.regular-expressions.info/posixbrackets.html

Before 5.18, the vertical tab (U+000B) wasn't included in \s.

$ diff -u <( unichars -au '\s' ) <( unichars -au '[[:space:]]' ) \
    && echo 'no difference'
--- /dev/fd/63  2013-05-21 22:08:03.000000000 -0400
+++ /dev/fd/62  2013-05-21 22:08:03.000000000 -0400
@@ -1,5 +1,6 @@
  ---- U+00009 CHARACTER TABULATION
  ---- U+0000A LINE FEED (LF)
+ ---- U+0000B LINE TABULATION
  ---- U+0000C FORM FEED (FF)
  ---- U+0000D CARRIAGE RETURN (CR)
  ---- U+00020 SPACE

`\s` also doesn't always match U+00A0, the non-breaking space. `/u` makes sure it does. `/a` makes sure it doesn't. — ikegami, May 22 '13 at 02:23

score 4 · Accepted Answer · answered May 22 '13 at 01:25

4

This is a POSIX character class, in this case a Unicode-friendly way of representing "any whitespace character".

See this page, scroll down to "POSIX Character Classes".

answered May 22 '13 at 01:25

Platinum Azure

45,269
12
110
134

POSIX predates Unicode by a large margin. The rationale for abstract character classes (my ad-hoc term, don't remember the official terminology) was more generally related to portability across locales and character sets. – tripleee May 22 '13 at 05:58

tadman · Answer 3 · 2013-05-22T05:14:32.453

0

There are a number of ways of expressing things like "whitespace character", and this is one of them. The set [...] allows the inclusion of things like [:space:] to add space characters to the set.

This reads as:

^ # At the beginning of string...
[[:space:]]* # ...zero or more whitespace characters...
@ # ...followed by an at sign.

edited May 22 '13 at 05:14

answered May 22 '13 at 01:26

tadman

208,517
23
234
262

Not space characters, *whitespace* characters. Maybe that's what you meant, but accurate terminology is more important than usual in this case, because many regex beginners start out believing that `\s` and/or `[[:space:]]` *do* match only the space character (`U+0020`). – Alan Moore May 22 '13 at 04:23
Whitespace is a better term, adjusted accordingly. – tadman May 22 '13 at 05:14

please explain this regular expression

3 Answers3