uri.ts

/*---------------------------------------------------------------------------------------------
 *  Copyright (c) Microsoft Corporation. All rights reserved.
 *  Licensed under the MIT License. See License.txt in the project root for license information.
 *--------------------------------------------------------------------------------------------*/
'use strict';

import { isWindows } from 'vs/base/common/platform';
import { CharCode } from 'vs/base/common/charCode';

const _schemePattern = /^\w[\w\d+.-]*$/;
const _singleSlashStart = /^\//;
const _doubleSlashStart = /^\/\//;

function _validateUri(ret: URI): void {
	// scheme, https://tools.ietf.org/html/rfc3986#section-3.1
	// ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
	if (ret.scheme && !_schemePattern.test(ret.scheme)) {
		throw new Error('[UriError]: Scheme contains illegal characters.');
	}

	// path, http://tools.ietf.org/html/rfc3986#section-3.3
	// If a URI contains an authority component, then the path component
	// must either be empty or begin with a slash ("/") character.  If a URI
	// does not contain an authority component, then the path cannot begin
	// with two slash characters ("//").
	if (ret.path) {
		if (ret.authority) {
			if (!_singleSlashStart.test(ret.path)) {
				throw new Error('[UriError]: If a URI contains an authority component, then the path component must either be empty or begin with a slash ("/") character');
			}
		} else {
			if (_doubleSlashStart.test(ret.path)) {
				throw new Error('[UriError]: If a URI does not contain an authority component, then the path cannot begin with two slash characters ("//")');
			}
		}
	}
}

// implements a bit of https://tools.ietf.org/html/rfc3986#section-5
function _referenceResolution(scheme: string, path: string): string {

	// the slash-character is our 'default base' as we don't
	// support constructing URIs relative to other URIs. This
	// also means that we alter and potentially break paths.
	// see https://tools.ietf.org/html/rfc3986#section-5.1.4
	switch (scheme) {
		case 'https':
		case 'http':
		case 'file':
			if (!path) {
				path = _slash;
			} else if (path[0] !== _slash) {
				path = _slash + path;
			}
			break;
	}
	return path;
}

const _empty = '';
const _slash = '/';
const _regexp = /^(([^:/?#]+?):)?(\/\/([^/?#]*))?([^?#]*)(\?([^#]*))?(#(.*))?/;

/**
 * Uniform Resource Identifier (URI) http://tools.ietf.org/html/rfc3986.
 * This class is a simple parser which creates the basic component paths
 * (http://tools.ietf.org/html/rfc3986#section-3) with minimal validation
 * and encoding.
 *
 *       foo://example.com:8042/over/there?name=ferret#nose
 *       \_/   \______________/\_________/ \_________/ \__/
 *        |           |            |            |        |
 *     scheme     authority       path        query   fragment
 *        |   _____________________|__
 *       / \ /                        \
 *       urn:example:animal:ferret:nose
 *
 *
 */
export default class URI implements UriComponents {

	static isUri(thing: any): thing is URI {
		if (thing instanceof URI) {
			return true;
		}
		if (!thing) {
			return false;
		}
		return typeof (<URI>thing).authority === 'string'
			&& typeof (<URI>thing).fragment === 'string'
			&& typeof (<URI>thing).path === 'string'
			&& typeof (<URI>thing).query === 'string'
			&& typeof (<URI>thing).scheme === 'string';
	}

	/**
	 * scheme is the 'http' part of 'http://www.msft.com/some/path?query#fragment'.
	 * The part before the first colon.
	 */
	readonly scheme: string;

	/**
	 * authority is the 'www.msft.com' part of 'http://www.msft.com/some/path?query#fragment'.
	 * The part between the first double slashes and the next slash.
	 */
	readonly authority: string;

	/**
	 * path is the '/some/path' part of 'http://www.msft.com/some/path?query#fragment'.
	 */
	readonly path: string;

	/**
	 * query is the 'query' part of 'http://www.msft.com/some/path?query#fragment'.
	 */
	readonly query: string;

	/**
	 * fragment is the 'fragment' part of 'http://www.msft.com/some/path?query#fragment'.
	 */
	readonly fragment: string;

	/**
	 * @internal
	 */
	protected constructor(scheme: string, authority: string, path: string, query: string, fragment: string);

	/**
	 * @internal
	 */
	protected constructor(components: UriComponents);

	/**
	 * @internal
	 */
	protected constructor(schemeOrData: string | UriComponents, authority?: string, path?: string, query?: string, fragment?: string) {

		if (typeof schemeOrData === 'object') {
			this.scheme = schemeOrData.scheme || _empty;
			this.authority = schemeOrData.authority || _empty;
			this.path = schemeOrData.path || _empty;
			this.query = schemeOrData.query || _empty;
			this.fragment = schemeOrData.fragment || _empty;
			// no validation because it's this URI
			// that creates uri components.
			// _validateUri(this);
		} else {
			this.scheme = schemeOrData || _empty;
			this.authority = authority || _empty;
			this.path = _referenceResolution(this.scheme, path || _empty);
			this.query = query || _empty;
			this.fragment = fragment || _empty;

			_validateUri(this);
		}
	}

	// ---- filesystem path -----------------------

	/**
	 * Returns a string representing the corresponding file system path of this URI.
	 * Will handle UNC paths and normalize windows drive letters to lower-case. Also
	 * uses the platform specific path separator. Will *not* validate the path for
	 * invalid characters and semantics. Will *not* look at the scheme of this URI.
	 */
	get fsPath(): string {
		return _makeFsPath(this);
	}

	// ---- modify to new -------------------------

	public with(change: { scheme?: string; authority?: string; path?: string; query?: string; fragment?: string }): URI {

		if (!change) {
			return this;
		}

		let { scheme, authority, path, query, fragment } = change;
		if (scheme === void 0) {
			scheme = this.scheme;
		} else if (scheme === null) {
			scheme = _empty;
		}
		if (authority === void 0) {
			authority = this.authority;
		} else if (authority === null) {
			authority = _empty;
		}
		if (path === void 0) {
			path = this.path;
		} else if (path === null) {
			path = _empty;
		}
		if (query === void 0) {
			query = this.query;
		} else if (query === null) {
			query = _empty;
		}
		if (fragment === void 0) {
			fragment = this.fragment;
		} else if (fragment === null) {
			fragment = _empty;
		}

		if (scheme === this.scheme
			&& authority === this.authority
			&& path === this.path
			&& query === this.query
			&& fragment === this.fragment) {

			return this;
		}

		return new _URI(scheme, authority, path, query, fragment);
	}

	// ---- parse & validate ------------------------

	public static parse(value: string): URI {
		const match = _regexp.exec(value);
		if (!match) {
			return new _URI(_empty, _empty, _empty, _empty, _empty);
		}
		return new _URI(
			match[2] || _empty,
			decodeURIComponent(match[4] || _empty),
			decodeURIComponent(match[5] || _empty),
			decodeURIComponent(match[7] || _empty),
			decodeURIComponent(match[9] || _empty),
		);
	}

	public static file(path: string): URI {

		let authority = _empty;

		// normalize to fwd-slashes on windows,
		// on other systems bwd-slashes are valid
		// filename character, eg /f\oo/ba\r.txt
		if (isWindows) {
			path = path.replace(/\\/g, _slash);
		}

		// check for authority as used in UNC shares
		// or use the path as given
		if (path[0] === _slash && path[1] === _slash) {
			let idx = path.indexOf(_slash, 2);
			if (idx === -1) {
				authority = path.substring(2);
				path = _slash;
			} else {
				authority = path.substring(2, idx);
				path = path.substring(idx) || _slash;
			}
		}

		return new _URI('file', authority, path, _empty, _empty);
	}

	public static from(components: { scheme?: string; authority?: string; path?: string; query?: string; fragment?: string }): URI {
		return new _URI(
			components.scheme,
			components.authority,
			components.path,
			components.query,
			components.fragment,
		);
	}

	// ---- printing/externalize ---------------------------

	/**
	 *
	 * @param skipEncoding Do not encode the result, default is `false`
	 */
	public toString(skipEncoding: boolean = false): string {
		return _asFormatted(this, skipEncoding);
	}

	public toJSON(): object {
		return this;
	}

	static revive(data: UriComponents | any): URI {
		if (!data) {
			return data;
		} else if (data instanceof URI) {
			return data;
		} else {
			let result = new _URI(data);
			result._fsPath = (<UriState>data).fsPath;
			result._formatted = (<UriState>data).external;
			return result;
		}
	}
}

export interface UriComponents {
	scheme: string;
	authority: string;
	path: string;
	query: string;
	fragment: string;
}

interface UriState extends UriComponents {
	$mid: number;
	fsPath: string;
	external: string;
}


// tslint:disable-next-line:class-name
class _URI extends URI {

	_formatted: string = null;
	_fsPath: string = null;

	get fsPath(): string {
		if (!this._fsPath) {
			this._fsPath = _makeFsPath(this);
		}
		return this._fsPath;
	}

	public toString(skipEncoding: boolean = false): string {
		if (!skipEncoding) {
			if (!this._formatted) {
				this._formatted = _asFormatted(this, false);
			}
			return this._formatted;
		} else {
			// we don't cache that
			return _asFormatted(this, true);
		}
	}

	toJSON(): object {
		const res = <UriState>{
			$mid: 1
		};
		// cached state
		if (this._fsPath) {
			res.fsPath = this._fsPath;
		}
		if (this._formatted) {
			res.external = this._formatted;
		}
		// uri components
		if (this.path) {
			res.path = this.path;
		}
		if (this.scheme) {
			res.scheme = this.scheme;
		}
		if (this.authority) {
			res.authority = this.authority;
		}
		if (this.query) {
			res.query = this.query;
		}
		if (this.fragment) {
			res.fragment = this.fragment;
		}
		return res;
	}
}

// reserved characters: https://tools.ietf.org/html/rfc3986#section-2.2
const encodeTable: { [ch: number]: string } = {
	[CharCode.Colon]: '%3A', // gen-delims
	[CharCode.Slash]: '%2F',
	[CharCode.QuestionMark]: '%3F',
	[CharCode.Hash]: '%23',
	[CharCode.OpenSquareBracket]: '%5B',
	[CharCode.CloseSquareBracket]: '%5D',
	[CharCode.AtSign]: '%40',

	[CharCode.ExclamationMark]: '%21', // sub-delims
	[CharCode.DollarSign]: '%24',
	[CharCode.Ampersand]: '%26',
	[CharCode.SingleQuote]: '%27',
	[CharCode.OpenParen]: '%28',
	[CharCode.CloseParen]: '%29',
	[CharCode.Asterisk]: '%2A',
	[CharCode.Plus]: '%2B',
	[CharCode.Comma]: '%2C',
	[CharCode.Semicolon]: '%3B',
	[CharCode.Equals]: '%3D',

	[CharCode.Space]: '%20',
};

function encodeURIComponentFast(uriComponent: string, allowSlash: boolean): string {
	let res: string = undefined;
	let nativeEncodePos = -1;

	for (let pos = 0; pos < uriComponent.length; pos++) {
		let code = uriComponent.charCodeAt(pos);

		// unreserved characters: https://tools.ietf.org/html/rfc3986#section-2.3
		if (
			(code >= CharCode.a && code <= CharCode.z)
			|| (code >= CharCode.A && code <= CharCode.Z)
			|| (code >= CharCode.Digit0 && code <= CharCode.Digit9)
			|| code === CharCode.Dash
			|| code === CharCode.Period
			|| code === CharCode.Underline
			|| code === CharCode.Tilde
			|| (allowSlash && code === CharCode.Slash)
		) {
			// check if we are delaying native encode
			if (nativeEncodePos !== -1) {
				res += encodeURIComponent(uriComponent.substring(nativeEncodePos, pos));
				nativeEncodePos = -1;
			}
			// check if we write into a new string (by default we try to return the param)
			if (res !== undefined) {
				res += uriComponent.charAt(pos);
			}

		} else {
			// encoding needed, we need to allocate a new string
			if (res === undefined) {
				res = uriComponent.substr(0, pos);
			}

			// check with default table first
			let escaped = encodeTable[code];
			if (escaped !== undefined) {

				// check if we are delaying native encode
				if (nativeEncodePos !== -1) {
					res += encodeURIComponent(uriComponent.substring(nativeEncodePos, pos));
					nativeEncodePos = -1;
				}

				// append escaped variant to result
				res += escaped;

			} else if (nativeEncodePos === -1) {
				// use native encode only when needed
				nativeEncodePos = pos;
			}
		}
	}

	if (nativeEncodePos !== -1) {
		res += encodeURIComponent(uriComponent.substring(nativeEncodePos));
	}

	return res !== undefined ? res : uriComponent;
}

function encodeURIComponentMinimal(path: string): string {
	let res: string = undefined;
	for (let pos = 0; pos < path.length; pos++) {
		let code = path.charCodeAt(pos);
		if (code === CharCode.Hash || code === CharCode.QuestionMark) {
			if (res === undefined) {
				res = path.substr(0, pos);
			}
			res += encodeTable[code];
		} else {
			if (res !== undefined) {
				res += path[pos];
			}
		}
	}
	return res !== undefined ? res : path;
}

/**
 * Compute `fsPath` for the given uri
 * @param uri
 */
function _makeFsPath(uri: URI): string {

	let value: string;
	if (uri.authority && uri.path.length > 1 && uri.scheme === 'file') {
		// unc path: file://shares/c$/far/boo
		value = `//${uri.authority}${uri.path}`;
	} else if (
		uri.path.charCodeAt(0) === CharCode.Slash
		&& (uri.path.charCodeAt(1) >= CharCode.A && uri.path.charCodeAt(1) <= CharCode.Z || uri.path.charCodeAt(1) >= CharCode.a && uri.path.charCodeAt(1) <= CharCode.z)
		&& uri.path.charCodeAt(2) === CharCode.Colon
	) {
		// windows drive letter: file:///c:/far/boo
		value = uri.path[1].toLowerCase() + uri.path.substr(2);
	} else {
		// other path
		value = uri.path;
	}
	if (isWindows) {
		value = value.replace(/\//g, '\\');
	}
	return value;
}

/**
 * Create the external version of a uri
 */
function _asFormatted(uri: URI, skipEncoding: boolean): string {

	const encoder = !skipEncoding
		? encodeURIComponentFast
		: encodeURIComponentMinimal;

	let res = '';
	let { scheme, authority, path, query, fragment } = uri;
	if (scheme) {
		res += scheme;
		res += ':';
	}
	if (authority || scheme === 'file') {
		res += _slash;
		res += _slash;
	}
	if (authority) {
		let idx = authority.indexOf('@');
		if (idx !== -1) {
			// <user>@<auth>
			const userinfo = authority.substr(0, idx);
			authority = authority.substr(idx + 1);
			idx = userinfo.indexOf(':');
			if (idx === -1) {
				res += encoder(userinfo, false);
			} else {
				// <user>:<pass>@<auth>
				res += encoder(userinfo.substr(0, idx), false);
				res += ':';
				res += encoder(userinfo.substr(idx + 1), false);
			}
			res += '@';
		}
		authority = authority.toLowerCase();
		idx = authority.indexOf(':');
		if (idx === -1) {
			res += encoder(authority, false);
		} else {
			// <auth>:<port>
			res += encoder(authority.substr(0, idx), false);
			res += authority.substr(idx);
		}
	}
	if (path) {
		// lower-case windows drive letters in /C:/fff or C:/fff
		if (path.length >= 3 && path.charCodeAt(0) === CharCode.Slash && path.charCodeAt(2) === CharCode.Colon) {
			let code = path.charCodeAt(1);
			if (code >= CharCode.A && code <= CharCode.Z) {
				path = `/${String.fromCharCode(code + 32)}:${path.substr(3)}`; // "/c:".length === 3
			}
		} else if (path.length >= 2 && path.charCodeAt(1) === CharCode.Colon) {
			let code = path.charCodeAt(0);
			if (code >= CharCode.A && code <= CharCode.Z) {
				path = `${String.fromCharCode(code + 32)}:${path.substr(2)}`; // "/c:".length === 3
			}
		}
		// encode the rest of the path
		res += encoder(path, true);
	}
	if (query) {
		res += '?';
		res += encoder(query, false);
	}
	if (fragment) {
		res += '#';
		res += encoder(fragment, false);
	}
	return res;
}