webext-patterns 1.1.1 → 1.3.0

package/index.d.ts

@@ -2,3 +2,5 @@ export declare const patternValidationRegex: RegExp;
 export declare const allStarsRegex: RegExp;
 export declare const allUrlsRegex: RegExp;
 export declare function patternToRegex(...matchPatterns: readonly string[]): RegExp;
+export declare function globToRegex(...globs: readonly string[]): RegExp;
+export declare function excludeDuplicatePatterns(matchPatterns: readonly string[]): string[];

package/index.js

@@ -1,9 +1,12 @@
+import escapeStringRegexp from 'escape-string-regexp';
 // Copied from https://github.com/mozilla/gecko-dev/blob/073cc24f53d0cf31403121d768812146e597cc9d/toolkit/components/extensions/schemas/manifest.json#L487-L491
 export const patternValidationRegex = /^(https?|wss?|file|ftp|\*):\/\/(\*|\*\.[^*/]+|[^*/]+)\/.*$|^file:\/\/\/.*$|^resource:\/\/(\*|\*\.[^*/]+|[^*/]+)\/.*$|^about:/;
 const isFirefox = typeof navigator === 'object' && navigator.userAgent.includes('Firefox/');
-export const allStarsRegex = isFirefox ? /^(https?|wss?):[/][/][^/]+([/].*)?$/ : /^https?:[/][/][^/]+([/].*)?$/;
+export const allStarsRegex = isFirefox
+    ? /^(https?|wss?):[/][/][^/]+([/].*)?$/
+    : /^https?:[/][/][^/]+([/].*)?$/;
 export const allUrlsRegex = /^(https?|file|ftp):[/]+/;
-function getRawRegex(matchPattern) {
+function getRawPatternRegex(matchPattern) {
     if (!patternValidationRegex.test(matchPattern)) {
         throw new Error(matchPattern + ' is an invalid pattern, it must match ' + String(patternValidationRegex));
     }
@@ -33,5 +36,50 @@ export function patternToRegex(...matchPatterns) {
     if (matchPatterns.includes('*://*/*')) {
         return allStarsRegex;
     }
-    return new RegExp(matchPatterns.map(x => getRawRegex(x)).join('|'));
+    return new RegExp(matchPatterns.map(x => getRawPatternRegex(x)).join('|'));
+}
+// The parens are required by .split() to preserve the symbols
+const globSymbols = /([?*]+)/;
+function splitReplace(part, index) {
+    if (part === '') {
+        // Shortcut for speed
+        return '';
+    }
+    if (index % 2 === 0) {
+        // Raw text, escape it
+        return escapeStringRegexp(part);
+    }
+    // Else: Symbol
+    if (part.includes('*')) { // Can be more than one and it swallows surrounding question marks
+        return '.*';
+    }
+    return [...part].map(() => isFirefox ? '.' : '.?').join('');
+}
+function getRawGlobRegex(glob) {
+    const regexString = glob
+        .split(globSymbols)
+        // eslint-disable-next-line unicorn/no-array-callback-reference -- tis ok 🤫
+        .map(splitReplace)
+        .join('');
+    // Drop "start with anything" and "end with anything" sequences because they're the default for regex
+    return ('^' + regexString + '$')
+        .replace(/^[.][*]/, '')
+        .replace(/[.][*]$/, '')
+        .replace(/^[$]$/, '.+'); // Catch `*` and `*`
+}
+export function globToRegex(...globs) {
+    // No glob, match anything; `include_globs: []` is the default
+    if (globs.length === 0) {
+        return /.*/;
+    }
+    return new RegExp(globs.map(x => getRawGlobRegex(x)).join('|'));
+}
+export function excludeDuplicatePatterns(matchPatterns) {
+    if (matchPatterns.includes('<all_urls>')) {
+        return ['<all_urls>'];
+    }
+    if (matchPatterns.includes('*://*/*')) {
+        return ['*://*/*'];
+    }
+    return matchPatterns.filter(possibleSubset => !matchPatterns.some(possibleSuperset => possibleSubset !== possibleSuperset && patternToRegex(possibleSuperset).test(possibleSubset)));
 }

package/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "webext-patterns",
-	"version": "1.1.1",
-	"description": "Tool to convert the patterns of your WebExtension manifest to regex",
+	"version": "1.3.0",
+	"description": "Tool to convert the patterns and globs of your WebExtension manifest to regex",
 	"keywords": [
 		"browser",
 		"chrome",
@@ -9,10 +9,15 @@
 		"firefox",
 		"glob",
 		"permission",
+		"greasemonkey",
+		"user scripts",
+		"globs",
+		"userscript",
 		"match",
 		"webext"
 	],
 	"repository": "fregante/webext-patterns",
+	"funding": "https://github.com/sponsors/fregante",
 	"license": "MIT",
 	"author": "Federico Brigante <me@fregante.com> (https://fregante.com)",
 	"type": "module",
@@ -24,6 +29,8 @@
 	],
 	"scripts": {
 		"build": "tsc",
+		"fix": "xo --fix",
+		"lint": "xo",
 		"prepare": "tsc --sourceMap false",
 		"test": "tsc && ava && xo",
 		"watch": "tsc --watch"
@@ -37,13 +44,25 @@
 			"unicorn/better-regex": "off"
 		}
 	},
+	"dependencies": {
+		"escape-string-regexp": "^5.0.0"
+	},
 	"devDependencies": {
-		"@sindresorhus/tsconfig": "^1.0.2",
-		"@types/chrome": "0.0.148",
-		"ava": "^3.15.0",
-		"sinon": "^11.1.1",
-		"type-fest": "^1.2.2",
-		"typescript": "^4.3.5",
-		"xo": "^0.42.0"
+		"@sindresorhus/tsconfig": "^3.0.1",
+		"@types/chrome": "0.0.210",
+		"ava": "^5.1.1",
+		"sinon": "^15.0.1",
+		"type-fest": "^3.5.3",
+		"typescript": "^4.9.4",
+		"xo": "^0.53.1"
+	},
+	"webExt": {
+		"sourceDir": "demo-extension",
+		"run": {
+			"startUrl": [
+				"chrome://extensions/",
+				"https://example.com/"
+			]
+		}
 	}
 }

package/readme.md

@@ -3,7 +3,7 @@
 [badge-gzip]: https://img.shields.io/bundlephobia/minzip/webext-patterns.svg?label=gzipped
 [link-bundlephobia]: https://bundlephobia.com/result?p=webext-patterns
 
-> Tool to convert the [patterns](https://developer.chrome.com/extensions/match_patterns) and [globs](https://wiki.greasespot.net/Include_and_exclude_rules) of your WebExtension manifest to regex
+> Tool to convert the [patterns](https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/manifest.json/content_scripts#globs) of your WebExtension manifest to regex
 
 This might be incomplete. Please help me test it by adding more pattern and URLs to the [tests](./test.js).
 
@@ -26,13 +26,35 @@ import {patternToRegex} from 'webext-patterns';
 
 ```js
 patternToRegex('http://*/*');
-// Returns /^http:\/\/?.+\/.+$/
+// Returns /^http:[/][/][^/]+[/].+$/
+
+globToRegex('*.example.com');
+// Returns /\.example\.com$/
+
+excludeDuplicatePatterns(['https://*.google.com/*', 'https://google.com/*']);
+// Returns ['https://*.google.com/*']
+```
+
+> **Note**
+> Firefox and Chrome handle patterns very slighly differently. `webext-patterns` defaults to Chrome’s logic, but if it detects a Firefox userAgent it will produce a Firefox-compatible regex.
+
+## API
+
+#### patternToRegex(pattern1, pattern2, etc)
+
+Accepts any number of `string` arguments and returns a single regex to match all of them.
+
+[Match patterns](https://developer.chrome.com/extensions/match_patterns) are used in the manifest’s permissions and content scripts’ `matches` and `exclude_matches` array.
+
+```js
+patternToRegex('http://*/*');
+// Returns /^http:[/][/][^/]+[/].+$/
 
 const gmailRegex = patternToRegex('*://mail.google.com/*');
 gmailRegex.test('https://mail.google.com/a/b/c'); // -> true
 gmailRegex.test('https://photos.google.com/a/b/c'); // -> false
 
-// Also accepts an array of patterns and returns a single regex
+// Also accepts multiple patterns and returns a single regex
 const googleRegex = patternToRegex(
 	'https://google.com/*',
 	'https://google.it/*'
@@ -41,12 +63,43 @@ googleRegex.test('https://google.it/search'); // -> true
 googleRegex.test('https://google.de/search'); // -> false
 ```
 
-## API
-
-#### patternToRegex(pattern1, pattern2, etc)
+#### globToRegex(pattern1, pattern2, etc)
 
 Accepts any number of `string` arguments and returns a single regex to match all of them.
 
+[Globs](https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/manifest.json/content_scripts#globs) are used in the manifest’s content scripts’ `include_globs` and `exclude_globs` arrays.
+
+```js
+globToRegex('*.example.co?');
+// Returns /\.example\.co.?$/ in Firefox
+// Returns /\.example\.co.$/ everywhere else
+
+const gmailRegex = globToRegex('*://mai?.google.com/*');
+gmailRegex.test('https://mail.google.com/a/b/c'); // -> true
+gmailRegex.test('https://photos.google.com/a/b/c'); // -> false
+
+// Also accepts multiple globs and returns a single regex
+const googleRegex = globToRegex(
+	'*google.com*',
+	'*google.it*'
+);
+googleRegex.test('https://google.it/search'); // -> true
+googleRegex.test('https://google.de/search'); // -> false
+```
+
+#### excludeDuplicatePatterns([pattern1, pattern2, etc])
+
+Accepts an array of patterns and returns a filtered array without the patterns that are already covered by others. For example `"https://*/*"` already covers all "https" URLs, so having `"https://google.com/*"` in the array won't make any difference and therefore it's dropped.
+
+```js
+excludeDuplicatePatterns([
+	"https://*/*",
+	"https://google.com/*",
+	"https://*.example.com/*",
+]);
+// Returns ["https://*/*"]
+```
+
 ## Related
 
 ### Permissions