Regex alternatives not cached in group

Question

How to catch alternative into the nested group?

I want to return first capturing group of: (.*)((core\/)?\/misc\/drupal\.js) (i.e. anything before the string core/misc/drupal.js which may or might not contain core/).

However, the core/ part always gets included to the first capturing group and not the second as supposed.

See also this for an example.

Which regex engine are you using? Different regex engines perform differently. — Patrick White, Jun 23 '16 at 15:00
I use the online tool for testing and intend to use `sed`in the bash script. — smihael, Jun 23 '16 at 15:02

score 1 · Accepted Answer · answered Jun 23 '16 at 15:06

1

There is a problem in your regex : you have one slash character at the end of the (core\/)? group, followed by another slash character in \/misc\/drupal\.js. This causes the core/ part to be parsed in the first dot rule no matter what.

Removing the additional slash is not enough though. To avoid the dot eating up as many characters as possible, you have to disable the default greedy behaviour by adding a ? to the * quantifier, like so :

(.*?)((core\/)?misc\/drupal\.js)

See it in action here.

answered Jun 23 '16 at 15:06

ttzn

2,543
22
26

1

Thanks! I didn't know about greediness of the Kleene star. Dies the same go for +? – smihael Jun 23 '16 at 15:12
@smihael absolutely. It also works for `?`. – ttzn Jun 23 '16 at 15:16

Regex alternatives not cached in group

1 Answers1