CODEWITHSUNDEEP
javascriptregex
guest271314
guest271314

Reputation: 1

Escaping characters within RegExp lookbehind assertions

At Regex to match url segment only if not preceded by specific parent segment

given the input

http://example.com/my-slug or http://example.com/my-slug/

the requirement is to match

"my-slug"

but not to match

"my-slug"

if preceded by

"news-events"

Was able to match "/my-slug" or "/my-slug/" by using RegExp lookbehind assertions at Chromium with --harmony flag set with RegExp

// why is the `\/` necessary following lookbehind?
let re = /(?<!news-events)\/[^/]+(?=\/$|$)/;

let sources = [
  "http://example.com/my-slug"
, "http://example.com/my-slug/"
, "http://example.com/news-events/my-slug"
];

for (let src of sources) {
  console.log(src.match(re))
}

however, when trying to precisely match "my-slug" without preceding "/" character by escaping the trailing "/" at news-event/ the RegExp does not return same match, where the match for "http://example.com/news-events/my-slug" is "y-slug" when the expected result is null

let re = /(?<!news-events\/)[^/]+(?=\/$|$)/;

let sources = [
  "http://example.com/my-slug"
, "http://example.com/my-slug/"
, "http://example.com/news-events/my-slug"
];

for (let src of sources) {
  console.log(src.match(re))
}

Questions:

Upvotes: 1

Views: 656

Answers (1)

Ry-
Ry-

Reputation: 224962

Since this bit of the regular expression:

[^/]+(?=\/$|$)

matches the text y-slug (y-slug is multiple non-slash characters, optionally followed by a slash, followed by the end of the string), and y-slug is not preceded by news-events/, it’s a valid match. Since my-slug doesn’t match, it’s also the first valid match, so it’s the match that is returned.

You could add a second, positive lookbehind to indicate that any match must be a complete segment.

let re = /(?<!news-events\/)(?<=\/)[^/]+(?=\/$|$)/;

let sources = [
  "http://example.com/my-slug"
, "http://example.com/my-slug/"
, "http://example.com/news-events/my-slug"
];

for (let src of sources) {
  console.log(src.match(re))
}

Upvotes: 1

Related Questions