Why is this recursive regex not working?

Question

I'm trying to write a simple regex to match nested curly brackets. So if I have this text:

{
  apple
  {second}
  banana
}

Then I want it to match the entire text between the first and last {} (including the 2nd pair of {}). Here's the regex I've written:

/{ (?:.+?|(?R) ) }/six

The output for this is:

{ apple {second}

As you can see the first curly bracket is being matched, and the 'banana' at the end is not being matched. Here's the output I want it to return:

apple {second} banana

What am I doing wrong?

Are you trying to match the entire inside, or do you want to match the inside as 3 separate tokens? — Corbin
– Corbin, Commented Nov 18, 2012 at 2:08
Seems like you don't need regex then? Just strip off the first/last characters. Or am I missing something? If you needed to make sure the braces were there, you could check the first and last characters explicitly, or use a simple pattern like {(.+)} — Corbin
– Corbin, Commented Nov 18, 2012 at 2:12
@Corbin Yes, you are missing something. The string I've posted is just a part of a larger text, so I need to use regex to parse it out. — Ali
– Ali, Commented Nov 18, 2012 at 2:12

mario · Accepted Answer · 2012-11-18 02:32:08Z

2

The pattern you want to use is:

/{ (?:  (?R) | .+? )+ }/six

With your regex, the .+? would have always taken precedence. PCRE would match the longest possible string and never look for the alternative.

Only making the alternative (..)+ repetetive allows the matching to switch between the recursive part and the match-anything placeholder.

answered Nov 18, 2012 at 2:32

mario

146k20 gold badges243 silver badges293 bronze badges

Sign up to request clarification or add additional context in comments.

8 Comments

Ali Over a year ago

Have you tried running this regex? When I run it as is, my apache always crashes.

Ali Over a year ago

What's the + after the last ) for? It seems to be the culprit and crashes apache, when I remove it then the regex doesn't capture correctly{}, but doesn't crash.

mario Over a year ago

Backtracking limit. Just tested on CLI. You might try to use [^{}]+ in place of the .+? though.

Ali Over a year ago

That worked. But its matching the full text instead of just the stuff inside the first set of brackets. Output: { apple {second} banana }

mario Over a year ago

That's what it is supposed to do. It always returns the complete match; only asserts the structure within, no cutting out.

|

John Kurlak · Accepted Answer · 2012-11-18 02:13:35Z

0

Does this work for you?

\{([\s\S]*)\}

answered Nov 18, 2012 at 2:13

John Kurlak

6,7407 gold badges45 silver badges59 bronze badges

1 Comment

Ali Over a year ago

Yes, but I prefer to do it via (?R) as I'm trying to learn how that works too. Can you tell me what's wrong with my earlier regex?

Collectives™ on Stack Overflow

Why is this recursive regex not working?

2 Answers 2

8 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

8 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related