Different and same Regex in Ruby and JS on the same input

Question

This is the input line "!!!??"

FIRST regex works similar in Ruby and JS and do what it need to do split this input for two lines contains "!" and "?" respectively:

Example #1 ruby like js /!+|\?+/g

RUBY s.scan(/!+|\?+/).inspect works like as JS s.match(/!+|\?+/g). And output is: RUBY [ '!!!', '??' ], JS [\"!!!\", \"??\"]

Example #2 ruby is not js /([?!])\1*/

Here ruby and js have different behavior RUBY s.scan(/([?!])\1*/) not equal to JS s.match(/([?!])\1*/g). RUBY returns two arrays [[\"!\"], [\"?\"]]. JS returns two strings as like as in the Example #1 [ '!!!', '??' ].

Why /([?!])\1*/ acts different in Ruby and JS?

@WiktorStribiżew: I don't know ruby, I just seen this and thought it looked correct, but on second look it seems you are correct — musefan
– musefan, Commented Nov 29, 2017 at 9:33

Wiktor Stribiżew · Accepted Answer · 2017-11-29 09:42:54Z

2

Since scan only returns captured substrings when capturing group(s) is/are defined in the pattern, you should modify the pattern to capture the whole match and add some more Ruby code (edited as per this comment):

s="!!!??"
matches = s.scan(/(([?!])\2*)/).inject([]) { |acc, (m, _)| acc << m }
puts matches
# = ['!!!', '??']

See the online Ruby demo.

As per @mudasobwa's comment, you may even contract that to

"!!!??".scan(/(([?!])\2*)/).map(&:first)

Here, (([?!])\2*) matches the same texts as /([?!])\1*/, but since the whole pattern is wrapped with capturing parentheses, the backreference to the ? or ! now has ID=2, hence \1 turns into \2. Inside the block, there is access to both the captured values via m (the whole match) and n (? or !). We are only collecting whole matches, hence m is added to matches upon each match.

edited Nov 29, 2017 at 9:42

answered Nov 29, 2017 at 9:14

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

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

4 Comments

Wiktor Stribiżew Over a year ago

@mudasobwa Yeah, this shorthand will make it shorter. Just in case anyone wants to have some "technical" access to the second capturing group, the expanded version might still be of value.

Aleksei Matiushkin Over a year ago

The expanded version should be "!!!??".scan(/(([?!])\2*)/).inject({}) { |acc, (m, _)| acc << m } without leaking an accumulator to the global space with a local variable :)

Wiktor Stribiżew Over a year ago

@mudasobwa That is interesting. I followed this answer when posting, and I bet you could say the same about that one, right?

Aleksei Matiushkin Over a year ago

Yes. And we can summon @Stefan to confirm :)

Collectives™ on Stack Overflow

Different and same Regex in Ruby and JS on the same input

1 Answer 1

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related