Javascript engine fails while executing this regex if given long strings to test

Question

This regex is written to disallow the urls starting with any url scheme and slashes(forward slash, backward slash) but will allow urls like "domain.tld" which are not starting with any url scheme or slashes. It should also allow the strings which are not url("some random input").

^(?!://)((?!//))(?!(.*?)*://)(?!:\\\\)(?!:/\\\\/\\\\)(?!(.*?)*:/\\\\/\\\\)(?!/\\\\/\\\\)(?!\\\\)(?!(.*?)*:\\\\)(?!www.)(?!(.*?)*.www.).*$

This regex works fine in java but in javscript, it is failing for longer strings.

Example: It works fine for "hey. hey hey hey hey" but starts taking time with "hey. hey hey hey hey " and hangs after "hey. hey hey hey hey hey hey"

Following are the cases which should be tested against the regex:

String                  | Expected result
__________________________________________
http://www.google.com   | False
HTTP://WWW.google.com   | False
adasd://www.google.com  | False
ftp://www.google.com    | False
mailto://www.google.com | False
//www.google.com        | False
://www.google.com       | False
www.google.com          | False
WWW.google.com          | False
test .http://google.com | False
skksdwww.google.com     | False
wWW.google.com          | False
://google.com           | False
.www.google.com         | False
as;;; .wwW.google.com   | False
as.wwW.google.com       | False
= #$@%@#.www.google.com | False
http:/\\/\\google.com   | False
:/\\/\\google.com       | False
http://gogle.com        | False
gogle.com //google.com  | False
google.com              | true
some random input       | true

What could be the problem in it?

UPDATE: I have updated the regex as per Wiktor Stribiżew's comment and it works fine.

See (.*?)* - this is a disastrous subpattern in larger patterns. Use a mere (.*?). Also, if you need to match a literal dot, escape it. See ^(?!://)((?!//))(?!(.*?)://)(?!:\\\\)(?!:/\\\\/\\\\)(?!(.*?):/\\\\/\\\\)(?!/\\\\/\\\\)(?!\\\\)(?!(.*?):\\\\)(?!www\.)(?!(.*?)\.www\.).*$. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Jul 21, 2016 at 10:24
I have to write it because I want to look for every possible url schemes(custom as well) — Neha Choudhary
– Neha Choudhary, Commented Jul 21, 2016 at 10:25
You do not have to say: match something unlimited number of times unlimited number of times. If you multiply infinity with infinity, what will you get? — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Jul 21, 2016 at 10:27
woah!! Thanks a TON!! works.. :) If you could write it as an answer I will accept it. — Neha Choudhary
– Neha Choudhary, Commented Jul 21, 2016 at 10:28
Well, I only showed where the issue is, but the pattern requires revamping for sure. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Jul 21, 2016 at 10:30

Wiktor Stribiżew · Accepted Answer · 2016-07-22 07:40:34Z

1

The (.*?)* subpattern is disastrous inside larger patterns. The nested * quantifiers (lazy inside and greedy outside) allow the regex engine to check a huge amount of substring variations before a failure occurs with a string that should not be matched.

Always test your patterns against strings that should not match.

Also, if you need to match a literal dot, escape it.

Here is your fixed and contracted regex:

^
  (?!.*?:?//)
  (?!:(?:/\\\\/)?\\\\)
  (?!(?:.*?:)?(?:/\\\\/)?\\\\)
  (?!(?:.*?\.)?www\.)
 .*
$

Or a one-liner:

^(?!.*?:?//)(?!:(?:/\\\\/)?\\\\)(?!(?:.*?:)?(?:/\\\\/)?\\\\)(?!(?:.*?\.)?www\.).*$

See the regex demo

answered Jul 22, 2016 at 7:40

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

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

Comments

Whothehellisthat · Accepted Answer · 2016-07-22 07:07:22Z

1

I didn't examine the whole thing (wow, that's a lot of slashes!) but you could greatly simplify the regex, I'm guessing. Just going from your post, maybe this would work for you:

/^(?!.*(?:www\.|\/\/|\/\\\\\/\\\\))/i

Should test negative for any protocol, including an empty one, including file: urls. Let me know if I've missed anything the regex needs to test for.

UPDATE: Now passes all tests.

edited Jul 22, 2016 at 7:07

answered Jul 21, 2016 at 10:29

Whothehellisthat

2,1521 gold badge15 silver badges15 bronze badges

3 Comments

Neha Choudhary Over a year ago

Na! doesnt work.. regEx.test("abc") // should return true, regEx.test("abc.com") // should return true

Whothehellisthat Over a year ago

Do you mean "abc" should return false? They both return true for me.

Neha Choudhary Over a year ago

ohh.. yeah. it works fine for both "abc" and "abc.com". checked with regex literal and wasn't working in that. But it doesn't solve every possible case. I will update my question with cases.

Collectives™ on Stack Overflow

Javascript engine fails while executing this regex if given long strings to test

2 Answers 2

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related