Avoid using RegExps in path to work around issue 19090.

pkg/path/lib/src/style/url.dart

 // Copyright (c) 2013, the Dart project authors.  Please see the AUTHORS file
 // for details. All rights reserved. Use of this source code is governed by a
 // BSD-style license that can be found in the LICENSE file.
 
 library path.style.url;
 
+import '../characters.dart' as chars;
 import '../internal_style.dart';
+import '../utils.dart';
 
 /// The style for URL paths.
 class UrlStyle extends InternalStyle {
   UrlStyle();
 
   final name = 'url';
   final separator = '/';
+  final separators = const ['/'];
+
+  // Deprecated properties.
+
   final separatorPattern = new RegExp(r'/');
   final needsSeparatorPattern = new RegExp(
       r"(^[a-zA-Z][-+.a-zA-Z\d]*://|[^/])$");
   final rootPattern = new RegExp(r"[a-zA-Z][-+.a-zA-Z\d]*://[^/]*");
   final relativeRootPattern = new RegExp(r"^/");
 
+  bool containsSeparator(String path) => path.contains('/');
+
+  bool isSeparator(int codeUnit) => codeUnit == chars.SLASH;
+
+  bool needsSeparator(String path) {
+    if (path.isEmpty) return false;
+
+    // A URL that doesn't end in "/" always needs a separator.
+    if (!isSeparator(path.codeUnitAt(path.length - 1))) return true;
+
+    // A URI that's just "scheme://" needs an extra separator, despite ending
+    // with "/".
+    var root = _getRoot(path);
+    return root != null && root.endsWith('://');
+  }
+
+  String getRoot(String path) {
+    var root = _getRoot(path);
+    return root == null ? getRelativeRoot(path) : root;
+  }
+
+  String getRelativeRoot(String path) {
+    if (path.isEmpty) return null;
+    return isSeparator(path.codeUnitAt(0)) ? "/" : null;
+  }
+
   String pathFromUri(Uri uri) => uri.toString();
 
   Uri relativePathToUri(String path) => Uri.parse(path);
   Uri absolutePathToUri(String path) => Uri.parse(path);
+
+  // A helper method for [getRoot] that doesn't handle relative roots.
+  String _getRoot(String path) {
+    if (path.isEmpty) return null;
+
+    // We aren't using a RegExp for this because they're slow (issue 19090). If
+    // we could, we'd match against r"[a-zA-Z][-+.a-zA-Z\d]*://[^/]*".
+
+    if (!isAlphabetic(path.codeUnitAt(0))) return null;
+    var start = 1;
+    for (; start < path.length; start++) {
+      var char = path.codeUnitAt(start);
+      if (isAlphabetic(char)) continue;
+      if (isNumeric(char)) continue;
+      if (char == chars.MINUS || char == chars.PLUS || char == chars.PERIOD) {
+        continue;
+      }
+
+      break;
+    }
+
+    if (start + 3 > path.length) return null;
+    if (path.substring(start, start + 3) != '://') return null;
+    start += 3;
+
+    // A URL root can end with a non-"/" prefix.
+    while (start < path.length && !isSeparator(path.codeUnitAt(start))) {
+      start++;
+    }
+    return path.substring(0, start);
+  }
 }