webext-patterns/distribution/index.js

Tool to convert the patterns and globs of your WebExtension manifest to regex

// Copied from https://github.com/mozilla/gecko-dev/blob/5836a062726f715fda621338a17b51aff30d0a8c/toolkit/components/extensions/schemas/manifest.json#L729-L741
export const patternValidationRegex = /^(https?|wss?|file|ftp|\*):\/\/(\*|\*\.[^*/]+|[^*/]+)\/.*$|^file:\/\/\/.*$|^resource:\/\/(\*|\*\.[^*/]+|[^*/]+)\/.*$|^about:/;
const isFirefox = globalThis.navigator?.userAgent.includes('Firefox/');
export const allStarsRegex = isFirefox
    ? /^(https?|wss?):[/][/][^/]+([/].*)?$/
    : /^https?:[/][/][^/]+([/].*)?$/;
export const allUrlsRegex = /^(https?|file|ftp):[/]+/;
export function assertValidPattern(matchPattern) {
    if (!isValidPattern(matchPattern)) {
        throw new Error(matchPattern + ' is an invalid pattern. See https://developer.mozilla.org/en-US/docs/Mozilla/Add-ons/WebExtensions/Match_patterns for more info.');
    }
}
export function isValidPattern(matchPattern) {
    return matchPattern === '<all_urls>' || patternValidationRegex.test(matchPattern);
}
export function testPatterns(url, patterns) {
    if (patterns.includes('<all_urls>') && allUrlsRegex.test(url)) {
        return true;
    }
    if (patterns.includes('*://*/*') && allStarsRegex.test(url)) {
        return true;
    }
    for (const pattern of patterns) {
        if (patternToRegex(pattern).test(url)) {
            return true;
        }
    }
    return false;
}
export function getMatchingPatterns(url, patterns) {
    return patterns.filter(pattern => testPatterns(url, [pattern]));
}
function getRawPatternRegex(matchPattern) {
    assertValidPattern(matchPattern);
    // Host undefined for file:///
    let [, protocol, host = '', pathname] = matchPattern.split(/(^[^:]+:[/][/])([^/]+)?/);
    protocol = protocol
        .replace('*', isFirefox ? '(https?|wss?)' : 'https?') // Protocol wildcard
        .replaceAll(/[/]/g, '[/]'); // Escape slashes
    if (host === '*') {
        host = '[^/]+';
    }
    host &&= host
        .replace(/^[*][.]/, '([^/]+.)*') // Initial wildcard
        .replaceAll(/[.]/g, '[.]') // Escape dots
        .replace(/[*]$/, '[^.]+'); // Last wildcard
    pathname = pathname
        .replaceAll(/[/]/g, '[/]') // Escape slashes
        .replaceAll(/[.]/g, '[.]') // Escape dots
        .replaceAll(/[*]/g, '.*'); // Any wildcard
    return '^' + protocol + host + '(' + pathname + ')?$';
}
export function patternToRegex(...matchPatterns) {
    // No pattern, match nothing https://stackoverflow.com/q/14115522/288906
    if (matchPatterns.length === 0) {
        return /$./;
    }
    if (matchPatterns.includes('<all_urls>')) {
        return allUrlsRegex;
    }
    if (matchPatterns.includes('*://*/*')) {
        return allStarsRegex;
    }
    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
        // eslint-disable-next-line no-use-extend-native/no-use-extend-native -- TODO: Drop after https://github.com/dustinspecker/eslint-plugin-no-use-extend-native/issues/157
        return RegExp.escape(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 removeRedundantPatterns(matchPatterns) {
    if (matchPatterns.includes('<all_urls>')) {
        return ['<all_urls>'];
    }
    if (matchPatterns.includes('*://*/*')) {
        return ['*://*/*'];
    }
    // Cover identical patterns
    const uniquePatterns = [...new Set(matchPatterns)];
    return uniquePatterns.filter(possibleSubset => 
    // Keep if there are no matches
    !uniquePatterns.some(possibleSuperset => 
    // Don't compare to self
    possibleSubset !== possibleSuperset
        // Drop if it's a subset
        && patternToRegex(possibleSuperset).test(possibleSubset)));
}