What tools are there for debugging/stepping through a regular expression?

Answer 1

Perhaps RegexCoach offers what you're looking for...

http://weitz.de/regex-coach/

Answer 2

That's probably so because it should not matter how the regex engine finds a match. The screenshot of RegexBuddy makes me want to use another regexp matcher, as there should not be a need to do backtracking.

Normally, you would translate the regular expression into an nondeterministic finite automaton (NFA) or a deterministic finite automaton (DFA), which can handle regular expressions without backtracking.

That being said, the only free tool for creating and testing regular expressions I know of is pyreb.

Answer 3

It's not quite as powerful as RegexBuddy, but it's a simple online interface: http://www.gskinner.com/RegExr/

You can mouse-over parts of your expression and it will tell you what it's doing. Very basic, but it can really save time when you do something stupid.

Answer 4

In Perl you can always just use re 'debug'; or use re 'debugcolor';

For example if you enter this into Perl:

use strict;
use warnings;
use 5.010;
use re 'debug';

# using the same strings as the question's image for reference:

my $str = 'Even if I do say so myself: "RegexBuddy is awesome"';
$str =~ /(Regexp?Buddy is (awful|acceptable|awesome))/;

This is what you get out:

Compiling REx "(Regexp?Buddy is (awful|acceptable|awesome))"
Final program:
   1: OPEN1 (3)
   3:   EXACT <Regex> (6)
   6:   CURLY {0,1} (10)
   8:     EXACT <p> (0)
  10:   EXACT <Buddy is > (14)
  14:   OPEN2 (16)
  16:     EXACT <a> (18)
  18:     TRIEC-EXACT[cw] (29) # this is a new feature in Perl 5.10
          <wful> 
          <cceptable> 
          <wesome> 
  29:   CLOSE2 (31)
  31: CLOSE1 (33)
  33: END (0)

anchored "Regex" at 0 floating "Buddy is a" at 5..6 (checking floating) minlen 19 
Guessing start of match in sv for REx "(Regexp?Buddy is (awful|acceptable|awesome))" against 'Even if I do say so myself: "RegexBuddy is awesome"'
Found floating substr "Buddy is a" at offset 34...
Found anchored substr "Regex" at offset 29...
Starting position does not contradict # /^/m...
Guessed: match at offset 29

Matching REx "(Regexp?Buddy is (awful|acceptable|awesome))" against 'RegexBuddy is awesome"'
  29 <'lf: "'> <RegexBuddy>  |  1:OPEN1(3)
  29 <'lf: "'> <RegexBuddy>  |  3:EXACT <Regex>(6)
  34 <Regex> <Buddy is a>    |  6:CURLY {0,1}(10)
                                  EXACT <p> can match 0 times out of 1...
  34 <Regex> <Buddy is a>    | 10:  EXACT <Buddy is >(14)
  43 <y is > <'awesome"'>    | 14:  OPEN2(16)
  43 <y is > <'awesome"'>    | 16:  EXACT <a>(18)
  44 < is a> <'wesome"'>     | 18:  TRIEC-EXACT[cw](29)
  44 < is a> <'wesome"'>     |      State:    2 Accepted:    0 Charid:  2 CP:  77 After State:    3
  45 < is aw> <'esome"'>     |      State:    3 Accepted:    0 Charid:  7 CP:  65 After State:   10
  46 < is awe> <'some"'>     |      State:   10 Accepted:    0 Charid:  b CP:  73 After State:   11
  47 < is awes> <'ome"'>     |      State:   11 Accepted:    0 Charid:  c CP:  6f After State:   12
  48 < is aweso> <'me"'>     |      State:   12 Accepted:    0 Charid:  d CP:  6d After State:   13
  49 < is awesom> <'e"'>     |      State:   13 Accepted:    0 Charid:  7 CP:  65 After State:   14
  50 < is awesome> <'"'>     |      State:   14 Accepted:    1 Charid:  3 CP:   0 After State:    0
                                    got 1 possible matches
                                    only one match left: #3 <wesome>
  50 < is awesome> <'"'>     | 29:  CLOSE2(31)
  50 < is awesome> <'"'>     | 31:  CLOSE1(33)
  50 < is awesome> <'"'>     | 33:  END(0)
Match successful!
Freeing REx: "(Regexp?Buddy is (awful|acceptable|awesome))"

( Note: I changed some parts of the output, so that it would highlight better )