use less regexp and more performance parsing

Author: sokra

lib/normalize.js

@@ -1,38 +1,86 @@
-var doubleSlashWinRegExp = /\\+/g;
-var doubleSlashNixRegExp = /\/+/g;
-var currentDirectoryWinMiddleRegExp = /\\(\.\\)+/;
-var currentDirectoryWinEndRegExp = /\\\.$/;
-var parentDirectoryWinMiddleRegExp = /\\+[^\\]+\\+\.\.\\/;
-var parentDirectoryWinEndRegExp1 = /([A-Z]:\\)\\*[^\\]+\\+\.\.$/i;
-var parentDirectoryWinEndRegExp2 = /\\+[^\\]+\\+\.\.$/;
-var currentDirectoryNixMiddleRegExp = /\/+(\.\/)+/;
-var currentDirectoryNixEndRegExp1 = /^\/+\.$/;
-var currentDirectoryNixEndRegExp2 = /\/+\.$/;
-var parentDirectoryNixMiddleRegExp = /(^|\/[^\/]+)\/+\.\.\/+/;
-var parentDirectoryNixEndRegExp1 = /^\/[^\/]+\/+\.\.$/;
-var parentDirectoryNixEndRegExp2 = /\/+[^\/]+\/+\.\.$/;
-var parentDirectoryNixEndRegExp3 = /^\/+\.\.$/;
-
-// RegExp magic :)
-
 module.exports = function normalize(path) {
-	while(currentDirectoryWinMiddleRegExp.test(path))
-		path = path.replace(currentDirectoryWinMiddleRegExp, "\\");
-	path = path.replace(currentDirectoryWinEndRegExp, "");
-	while(parentDirectoryWinMiddleRegExp.test(path))
-		path = path.replace(parentDirectoryWinMiddleRegExp, "\\");
-	path = path.replace(parentDirectoryWinEndRegExp1, "$1");
-	path = path.replace(parentDirectoryWinEndRegExp2, "");
-
-	while(currentDirectoryNixMiddleRegExp.test(path))
-		path = path.replace(currentDirectoryNixMiddleRegExp, "/");
-	path = path.replace(currentDirectoryNixEndRegExp1, "/");
-	path = path.replace(currentDirectoryNixEndRegExp2, "");
-	while(parentDirectoryNixMiddleRegExp.test(path))
-		path = path.replace(parentDirectoryNixMiddleRegExp, "/");
-	path = path.replace(parentDirectoryNixEndRegExp1, "/");
-	path = path.replace(parentDirectoryNixEndRegExp2, "");
-	path = path.replace(parentDirectoryNixEndRegExp3, "/");
-
-	return path.replace(doubleSlashWinRegExp, "\\").replace(doubleSlashNixRegExp, "/");
-};
+	var parts = path.split(/(\\+|\/+)/);
+	if(parts.length === 1)
+		return path;
+	var result = [];
+	var absolutePathStart = 0;
+	for(var i = 0, sep = false; i < parts.length; i++, sep = !sep) {
+		var part = parts[i];
+		if(i === 0 && /^([A-Z]:)?$/i.test(part)) {
+			result.push(part);
+			absolutePathStart = 2;
+		} else if(sep) {
+			result.push(part[0]);
+		} else if(part === "..") {
+			switch(result.length) {
+				case 0:
+					// i. e. ".." => ".."
+					// i. e. "../a/b/c" => "../a/b/c"
+					result.push(part);
+					break;
+				case 2:
+					// i. e. "a/.." => ""
+					// i. e. "/.." => "/"
+					// i. e. "C:\.." => "C:\"
+					// i. e. "a/../b/c" => "b/c"
+					// i. e. "/../b/c" => "/b/c"
+					// i. e. "C:\..\a\b\c" => "C:\a\b\c"
+					i++;
+					sep = !sep;
+					result.length = absolutePathStart;
+					break;
+				case 4:
+					// i. e. "a/b/.." => "a"
+					// i. e. "/a/.." => "/"
+					// i. e. "C:\a\.." => "C:\"
+					// i. e. "/a/../b/c" => "/b/c"
+					if(absolutePathStart === 0) {
+						result.length -= 3;
+					} else {
+						i++;
+						sep = !sep;
+						result.length = 2;
+					}
+					break;
+				default:
+					// i. e. "/a/b/.." => "/a"
+					// i. e. "/a/b/../c" => "/a/c"
+					result.length -= 3;
+					break;
+			}
+		} else if(part === ".") {
+			switch(result.length) {
+				case 0:
+					// i. e. "." => "."
+					// i. e. "./a/b/c" => "./a/b/c"
+					result.push(part);
+					break;
+				case 2:
+					// i. e. "a/." => "a"
+					// i. e. "/." => "/"
+					// i. e. "C:\." => "C:\"
+					// i. e. "C:\.\a\b\c" => "C:\a\b\c"
+					if(absolutePathStart === 0) {
+						result.length--;
+					} else {
+						i++;
+						sep = !sep;
+					}
+					break;
+				default:
+					// i. e. "a/b/." => "a/b"
+					// i. e. "/a/." => "/"
+					// i. e. "C:\a\." => "C:\"
+					// i. e. "a/./b/c" => "a/b/c"
+					// i. e. "/a/./b/c" => "/a/b/c"
+					result.length--;
+					break;
+			}
+		} else if(part) {
+			result.push(part);
+		}
+	}
+	if(result.length === 1 && /^[A-Za-z]:$/.test(result))
+		return result[0] + "\\";
+	return result.join("");
+};