feat: be possible to skip re2 (#656)

README.md

@@ -279,6 +279,15 @@ Default: `true`
 
 Ensure the URL provided is validated as a [WHATWG URL](https://nodejs.org/api/url.html#url_the_whatwg_url_api) API compliant.
 
+## Environment Variables
+
+#### METASCRAPER_RE2
+
+Type: `boolean`<br>
+Default: `true`
+
+It attemptt to load re2 to use instead of RegExp.
+
 ## Benchmark
 
 To give you an idea of how accurate **metascraper** is, here is a comparison of similar libraries:

packages/metascraper-helpers/index.js

@@ -1,7 +1,6 @@
 'use strict'
 
 const memoizeOne = require('memoize-one').default || require('memoize-one')
-const urlRegex = require('url-regex-safe')({ exact: true, parens: true })
 const condenseWhitespace = require('condense-whitespace')
 const { getExtension: mimeExtension } = require('mime')
 const capitalize = require('microsoft-capitalize')
@@ -19,6 +18,14 @@ const isUri = require('is-uri')
 const { URL } = require('url')
 const tldts = require('tldts')
 
+const urlRegex = require('url-regex-safe')({
+  exact: true,
+  parens: true,
+  re2: process.env.METASCRAPER_RE2
+    ? process.env.METASCRAPER_RE2 === 'true'
+    : undefined
+})
+
 const {
   chain,
   eq,

packages/metascraper-helpers/package.json

@@ -42,7 +42,7 @@
     "re2": "~1.20.1",
     "smartquotes": "~2.3.2",
     "tldts": "~6.0.12",
-    "url-regex-safe": "~3.0.0",
+    "url-regex-safe": "~4.0.0",
     "video-extensions": "~1.2.0"
   },
   "devDependencies": {