keep input (for toString) when using URI.parse, URI.file uses now URI.parse

Add tests that check our logic against Nodejs's pathToFileURL

src/vs/base/common/uri.ts

@@ -74,8 +74,17 @@ function _referenceResolution(scheme: string, path: string): string {
 
 const _empty = '';
 const _slash = '/';
+
 const _regexp = /^(([^:/?#]+?):)?(\/\/([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?/;
 
+const enum MatchIndex {
+	scheme = 2,
+	authority = 4,
+	path = 5,
+	query = 7,
+	fragment = 9
+}
+
 /**
  * Uniform Resource Identifier (URI) http://tools.ietf.org/html/rfc3986.
  * This class is a simple parser which creates the basic component parts
@@ -262,15 +271,24 @@ export class URI implements UriComponents {
 	static parse(value: string): URI {
 		const match = _regexp.exec(value);
 		if (!match) {
-			return new _URI(_empty, _empty, _empty, _empty, _empty);
+			throw new Error(`[UriError]: Invalid input: ${value}`);
 		}
-		return new _URI(
-			match[2] || _empty,
-			percentDecode(match[4] || _empty),
-			percentDecode(match[5] || _empty),
-			percentDecode(match[7] || _empty),
-			percentDecode(match[9] || _empty)
+
+		const scheme = _schemeFix(match[MatchIndex.scheme]) || _empty;
+		const authority = match[MatchIndex.authority] || _empty;
+		const path = _referenceResolution(scheme, match[MatchIndex.path] || _empty);
+		const query = match[MatchIndex.query] || _empty;
+		const fragment = match[MatchIndex.fragment] || _empty;
+
+		const result = new _URI(
+			scheme,
+			percentDecode(authority),
+			percentDecode(path),
+			percentDecode(query),
+			percentDecode(fragment),
 		);
+		result._formatted = _toString(false, scheme, authority, path, query, fragment);
+		return result;
 	}
 
 	/**
@@ -318,7 +336,15 @@ export class URI implements UriComponents {
 			}
 		}
 
-		return new _URI('file', authority, path, _empty, _empty);
+		if (path.charAt(0) !== _slash) {
+			path = _slash + path;
+		}
+
+		// escape some vital characters
+		authority = authority.replace(/%/g, '%25');
+		path = path.replace(/%/g, '%25');
+
+		return URI.parse('file://' + authority + path);
 	}
 
 	static from(components: { scheme: string; authority?: string; path?: string; query?: string; fragment?: string; }): URI {

src/vs/base/test/common/uri.test.ts → src/vs/base/test/node/uri.test.ts

@@ -5,7 +5,7 @@
 import * as assert from 'assert';
 import { URI, UriComponents } from 'vs/base/common/uri';
 import { isWindows } from 'vs/base/common/platform';
-
+import { pathToFileURL } from 'url';
 
 suite('URI', () => {
 	test('file#toString', () => {
@@ -427,21 +427,24 @@ suite('URI', () => {
 	});
 
 	test('Unable to open \'%A0.txt\': URI malformed #76506', function () {
-
-		let uri = URI.file('/foo/%A0.txt');
-		let uri2 = URI.parse(uri.toString());
-		assert.equal(uri.scheme, uri2.scheme);
-		assert.equal(uri.path, uri2.path);
-
-		uri = URI.file('/foo/%2e.txt');
-		uri2 = URI.parse(uri.toString());
-		assert.equal(uri.scheme, uri2.scheme);
-		assert.equal(uri.path, uri2.path);
+		let uriFromPath = URI.file('/foo/%A0.txt');
+		let uriFromStr = URI.parse(uriFromPath.toString());
+		assert.equal(uriFromPath.scheme, uriFromStr.scheme);
+		assert.equal(uriFromPath.path, uriFromStr.path);
+		assert.equal(uriFromPath.toString(), 'file:///foo/%25A0.txt');
+		assert.equal(uriFromStr.toString(), 'file:///foo/%25A0.txt');
 	});
 
+	test('Valid percent-encoded character in filename', function () {
+		let uriFromPath = URI.file('/foo/%2e.txt');
+		let uriFromStr = URI.parse(uriFromPath.toString());
+		assert.equal(uriFromPath.scheme, uriFromStr.scheme);
+		assert.equal(uriFromPath.path, uriFromStr.path);
+		assert.equal(uriFromPath.toString(), 'file:///foo/%252e.txt');
+		assert.equal(uriFromStr.toString(), 'file:///foo/%252e.txt');
+	});
 
 	test('Links in markdown are broken if url contains encoded parameters #79474', function () {
-		this.skip();
 		let strIn = 'https://myhost.com/Redirect?url=http%3A%2F%2Fwww.bing.com%3Fsearch%3Dtom';
 		let uri1 = URI.parse(strIn);
 		let strOut = uri1.toString();
@@ -456,7 +459,6 @@ suite('URI', () => {
 	});
 
 	test('Uri#parse can break path-component #45515', function () {
-		this.skip();
 		let strIn = 'https://firebasestorage.googleapis.com/v0/b/brewlangerie.appspot.com/o/products%2FzVNZkudXJyq8bPGTXUxx%2FBetterave-Sesame.jpg?alt=media&token=0b2310c4-3ea6-4207-bbde-9c3710ba0437';
 		let uri1 = URI.parse(strIn);
 		let strOut = uri1.toString();
@@ -499,4 +501,20 @@ suite('URI', () => {
 		// }
 		// console.profileEnd();
 	});
+
+	function assertFileUri(path: string): void {
+		// check that our uri aligns with nodejs
+		const actual = URI.file(path).toString();
+		const expected = pathToFileURL(path).href;
+		assert.equal(actual, expected);
+	}
+
+	test('URI.file vs pathToFileURL', function () {
+		assertFileUri('/foo/bar');
+		assertFileUri('/foo/%2e.txt'); // %2e -> .
+		assertFileUri('/foo/%A0.txt'); // %A0 -> invalid
+		assertFileUri('/foo/ü.txt');
+		assertFileUri('/foo/ß.txt');
+		// assertFileUri('foo'); nodejs resolves the path first
+	});
 });