fix: hexadecimal code-points are not converted (#2)

capybara1 · web-flow · commit 816790b58c68 · 2021-01-13T13:11:07.000+01:00
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -4,6 +4,12 @@ All notable changes to the "vscode-url-title-resolver" extension will be documen
 
 Check [Keep a Changelog](http://keepachangelog.com/) for recommendations on how to structure this file.
 
+## [1.1.1]
+
+### Changed
+
+- Fixed hexadecimal code-points are not converted
+
 ## [1.1.0]
 
 ### Added
diff --git a/package.json b/package.json
@@ -2,7 +2,7 @@
 	"name": "vscode-url-title-resolver",
 	"displayName": "URL title resolver for Markdown",
 	"description": "Resolves selected URLs of one or more HTML pages and uses the document title(s) to format Markdown links",
-	"version": "1.1.0",
+	"version": "1.1.1",
 	"publisher": "capybara1",
 	"repository": {
 		"type": "git",
diff --git a/src/lib/htmlHelper.ts b/src/lib/htmlHelper.ts
@@ -240,7 +240,7 @@ const htmlEntityMapping = new Map<string, string>([
 	['diams', '♦']
 ]);
 
-const htmlEntityRegExp = new RegExp(`&(#\\d+|${Array.from(htmlEntityMapping.keys()).join('|')});`, 'gi');
+const htmlEntityRegExp = new RegExp(`&(#\\d+|#x[\\da-fA-F]+|${Array.from(htmlEntityMapping.keys()).join('|')});`, 'gi');
 
 export function getTitle(html: string): string|null {
 	const matches = html.match(/(?<=\<title\>).*?(?=\<\/title\>)/si);
@@ -262,12 +262,16 @@ function sanitizeTitle(title: string): string {
 function resolveHtmlEntities(html: string) {
 	return html.replace(
 		htmlEntityRegExp,
-		(_, p1: string): string => {
+		(match, p1: string): string => {
 			const char = htmlEntityMapping.get(p1);
 			if (char) {
 				return char;
 			}
-			const codePoint = +p1.substr(1);
-			return String.fromCodePoint(codePoint);
+			const codePoint = +p1.substr(1) || parseInt(p1.substr(2), 16);
+			const result = codePoint
+				? String.fromCodePoint(codePoint)
+				: match;
+
+			return result;
 		});
-}
+}
diff --git a/src/test/suite/htmlHelper.test.ts b/src/test/suite/htmlHelper.test.ts
@@ -0,0 +1,61 @@
+import * as assert from 'assert';
+
+import { getTitle } from '../../lib/htmlHelper';
+
+const plainTitle = 'test';
+const toHtmlContent = (title: string) => `<html><head><title>${title}</title></head><body></body></html>`;
+
+suite('htmlHelper Test Suite', () => {
+
+	test('Given no title element, getTitle returns null', () => {
+		const input = '';
+		const result = getTitle(input);
+		assert.strictEqual(result, null);
+	});
+
+	test('Given a title element with plain title, getTitle returns the plain title', () => {
+		const input = toHtmlContent(plainTitle);
+		const result = getTitle(input);
+		assert.strictEqual(result, plainTitle);
+	});
+
+	test('Given a title element with leading whitespace, getTitle returns the plain title', () => {
+		const input = toHtmlContent(' \n\t ' + plainTitle);
+		const result = getTitle(input);
+		assert.strictEqual(result, plainTitle);
+	});
+
+	test('Given a title element with trailing whitespace, getTitle returns the plain title', () => {
+		const input = toHtmlContent(plainTitle + ' \n\t ');
+		const result = getTitle(input);
+		assert.strictEqual(result, plainTitle);
+	});
+
+	test('Given a title element with named entity, getTitle returns the title with resolved entity', () => {
+		const input = toHtmlContent('&uuml;');
+		const expected = 'ü';
+		const result = getTitle(input);
+		assert.strictEqual(result, expected);
+	});
+
+	test('Given a title element with decimal code point entity, getTitle returns the title with resolved entity', () => {
+		const input = toHtmlContent('&#252;');
+		const expected = 'ü';
+		const result = getTitle(input);
+		assert.strictEqual(result, expected);
+	});
+
+	test('Given a title element with hexadecimal code point entity, getTitle returns the title with resolved entity', () => {
+		const input = toHtmlContent('&#xFC;');
+		const expected = 'ü';
+		const result = getTitle(input);
+		assert.strictEqual(result, expected);
+	});
+
+	test('Given a title element with unknown entity, getTitle returns the title with unresolved entity', () => {
+		const input = toHtmlContent('&unknown;');
+		const expected = '&unknown;';
+		const result = getTitle(input);
+		assert.strictEqual(result, expected);
+	});
+});
