Further improve CJK and encoded URL support (#61)

Co-authored-by: Federico <[email protected]>

Author: scarf005

__snapshots__/index.test.js.snap

@@ -40,6 +40,8 @@ exports[`https://github.com/fregante/shorten-repo-url/blame/v0.12/.gitignore 1`]
 
 exports[`https://github.com/fregante/shorten-repo-url/blob/cc8fc46/.gitignore 1`] = `<code>cc8fc46</code>/.gitignore`;
 
+exports[`https://github.com/fregante/shorten-repo-url/blob/main/한글.txt 1`] = `<code>main</code>/한글.txt`;
+
 exports[`https://github.com/fregante/shorten-repo-url/blob/master/.gitignore 1`] = `<code>master</code>/.gitignore`;
 
 exports[`https://github.com/fregante/shorten-repo-url/blob/v0.12/.gitignore 1`] = `<code>v0.12</code>/.gitignore`;
@@ -202,6 +204,10 @@ exports[`https://github.com/refined-github/refined-github/wiki/%22Can-you-add-th
 
 exports[`https://github.com/refined-github/refined-github/wiki/%22Can-you-add-this-feature%3F%22#3-it-doesnt-require-options 1`] = `Wiki: "Can you add this feature?" (3 it doesnt require options) (refined-github/refined-github)`;
 
+exports[`https://github.com/scarf005/hangul-test/wiki/한글-위키-페이지 1`] = `Wiki: 한글 위키 페이지 (scarf005/hangul-test)`;
+
+exports[`https://github.com/scarf005/hangul-test/wiki/한글-위키-페이지#한글-헤딩 1`] = `Wiki: 한글 위키 페이지 (한글 헤딩) (scarf005/hangul-test)`;
+
 exports[`https://github.com/settings/profile 1`] = `github.com/settings/profile`;
 
 exports[`https://github.com/sindresorhus 1`] = `@sindresorhus`;
@@ -261,3 +267,5 @@ exports[`https://www.npmjs.com/ 1`] = `npmjs.com`;
 exports[`https://www.npmjs.com/packaasdge/node 1`] = `npmjs.com/packaasdge/node`;
 
 exports[`https://wwww.google.com/ 1`] = `wwww.google.com`;
+
+exports[`https://한글로-된-경로.com/하위경로#한글-해시 1`] = `한글로-된-경로.com/하위경로#한글-해시`;

index.js

@@ -81,16 +81,15 @@ function shortenRepoUrl(href, currentUrl = 'https://github.com') {
 	const currentRepo = currentUrl.pathname.slice(1).split('/', 2).join('/');
 
 	/**
-	 * Parse URL
+	 * Parse URL manually to avoid URL encoding and punycode
 	 */
+	const origin = href.split('/', 3).join('/');
+	const pathname = href.slice(origin.length).replace(/[?#].*/, '');
+	const hash = /#.+$/.exec(href)?.[0] ?? '';
+
+	// Use URL exclusively for search parameters because they're too hard to parse
 	const url = new URL(href);
-	const {
-		origin,
-		pathname,
-		search,
-		searchParams,
-		hash,
-	} = url;
+	const {search, searchParams} = url;
 
 	const pathnameParts = pathname.slice(1).split('/'); // ['user', 'repo', 'pull', '342']
 	const repoPath = pathnameParts.slice(2).join('/'); // 'pull/342'
@@ -158,31 +157,24 @@ function shortenRepoUrl(href, currentUrl = 'https://github.com') {
 	 * Shorten URL
 	 */
 	if (isReserved || pathname === '/' || (!isLocal && !isRaw && !isRedirection)) {
-		const parsedUrl = new URL(href);
 		const cleanHref = [
-			parsedUrl.origin
+			origin
 				.replace(/^https:[/][/]/, '')
 				.replace(/^www[.]/, ''),
-			parsedUrl.pathname
+			pathname
 				.replace(/[/]$/, ''),
 		];
 
 		if (['issues', 'pulls'].includes(user) && !repo) {
-			const query = pullQueryOut(parsedUrl.searchParams, parsedUrl.pathname);
-			cleanHref.push(parsedUrl.search, query);
+			const query = pullQueryOut(url.searchParams, url.pathname);
+			cleanHref.push(url.search, query);
 		} else {
-			cleanHref.push(parsedUrl.search);
+			cleanHref.push(url.search);
 		}
 
-		cleanHref.push(parsedUrl.hash);
+		cleanHref.push(decodeURI(url.hash));
 
-		// The user prefers seeing the URL as it was typed, so we need to decode it
-		try {
-			return decodeURI(cleanHref.join(''));
-		} catch {
-			// Decoding fails if the URL includes '%%'
-			return href;
-		}
+		return cleanHref.join('');
 	}
 
 	if (user && !repo) {
@@ -290,16 +282,18 @@ function shortenRepoUrl(href, currentUrl = 'https://github.com') {
  */
 function safeDecode(url) {
 	try {
-		return decodeURIComponent(url);
+		return new URL(url).href;
 	} catch {
 		return url;
 	}
 }
 
+/** @param a {HTMLAnchorElement} */
 function getLinkHref(a) {
 	return a.dataset.originalHref ?? a.href;
 }
 
+/** @param a {HTMLAnchorElement} */
 function isCustomLink(a) {
 	const url = safeDecode(getLinkHref(a));
 	const label = safeDecode(a.textContent);
@@ -321,7 +315,7 @@ export function applyToLink(a, currentUrl) {
 		// And if there are no additional images in the link
 		&& !a.firstElementChild
 	) {
-		const url = getLinkHref(a);
+		const url = a.textContent;
 		const shortened = shortenRepoUrl(url, currentUrl);
 		a.replaceChildren(
 			...shortened.split(

index.test.js

@@ -32,6 +32,7 @@ const urls = [
 	'https://github.com/fregante/shorten-repo-url/blob/master/.gitignore',
 	'https://github.com/fregante/shorten-repo-url/blob/v0.12/.gitignore',
 	'https://github.com/fregante/shorten-repo-url/blob/cc8fc46/.gitignore',
+	'https://github.com/fregante/shorten-repo-url/blob/main/한글.txt',
 	'https://github.com/nodejs/node/blob/master/.gitignore',
 	'https://github.com/nodejs/node/blob/v0.12/.gitignore',
 	'https://github.com/nodejs/node/blob/cc8fc46/.gitignore',
@@ -138,6 +139,8 @@ const urls = [
 	'https://github.com/refined-github/refined-github/wiki/%22Can-you-add-this-feature%3F%22#',
 	'https://github.com/refined-github/refined-github/wiki/%22Can-you-add-this-feature%3F%22',
 	'https://github.com/fregante/shorten-repo-url/wiki/%22Can-you-add-this-feature%3F%22',
+	'https://github.com/scarf005/hangul-test/wiki/한글-위키-페이지',
+	'https://github.com/scarf005/hangul-test/wiki/한글-위키-페이지#한글-헤딩',
 	'https://developer.mozilla.org/en-US/docs/Web/API/Document/createElement#parameters',
 	'https://www.google.com/',
 	'https://wwww.google.com/',
@@ -146,6 +149,7 @@ const urls = [
 	'https://www.npmjs.com/packaasdge/node',
 	'https://example.com/nodejs/node/blob/cc8fc46/.gitignore',
 	'https://example.site/한글로-된-URL',
+	'https://한글로-된-경로.com/하위경로#한글-해시',
 ];
 
 test.each(urls)('%s', url => {