big update to observe, template_binding, polymer

pkg/template_binding/lib/src/mustache_tokens.dart

Index: pkg/template_binding/lib/src/mustache_tokens.dart
diff --git a/pkg/template_binding/lib/src/mustache_tokens.dart b/pkg/template_binding/lib/src/mustache_tokens.dart
new file mode 100644
index 0000000000000000000000000000000000000000..ece3b60bc5af64fec58874df20e0ecfe3def1147
--- /dev/null
+++ b/pkg/template_binding/lib/src/mustache_tokens.dart
@@ -0,0 +1,143 @@
+// 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 template_binding.src.mustache_tokens;
+
+import 'package:observe/observe.dart';
+import 'package:template_binding/template_binding.dart';
+
+/**
+ * Represents a set of parsed tokens from a {{ mustache binding expression }}.
+ * This can be created by calling [parse].
+ *
+ * For performance reasons the data is stored in one linear array in [_tokens].
+ * This class wraps that array and provides accessors in an attempt to make the
+ * pattern easier to understand. See [length] and [getText] for example.
+ */
+class MustacheTokens {
+  // Constants for indexing into the exploded structs in [_tokens] .
+  static const _TOKEN_TEXT = 0;
+  static const _TOKEN_ONETIME = 1;
+  static const _TOKEN_PATH = 2;
+  static const _TOKEN_PREPAREFN = 3;
+  static const _TOKEN_SIZE = 4;
+
+  // There is 1 extra entry for the end text.
+  static const _TOKEN_ENDTEXT = 1;
+
+  bool get hasOnePath => _tokens.length == _TOKEN_SIZE + _TOKEN_ENDTEXT;
+  bool get isSimplePath => hasOnePath &&
+      _tokens[_TOKEN_TEXT] == '' && _tokens[_TOKEN_SIZE + _TOKEN_TEXT] == '';
+
+  /**
+   * [TEXT, (ONE_TIME?, PATH, DELEGATE_FN, TEXT)+] if there is at least one
+   * mustache.
+   */
+  final List _tokens;
+
+  final bool onlyOneTime;
+
+  // Dart note: I think this is cached in JavaScript to avoid an extra
+  // allocation per template instance. Seems reasonable, so we do the same.
+  Function _combinator;
+  Function get combinator => _combinator;
+
+  MustacheTokens._(this._tokens, this.onlyOneTime) {
+    // Should be: [TEXT, (ONE_TIME?, PATH, DELEGATE_FN, TEXT)+].
+    assert((_tokens.length - _TOKEN_ENDTEXT) % _TOKEN_SIZE == 0);
+
+    _combinator = hasOnePath ? _singleCombinator : _listCombinator;
+  }
+
+  int get length => _tokens.length ~/ _TOKEN_SIZE;
+
+  /**

[justinfagnani, 2014/02/04 23:34:49]

thanks!!

[Jennifer Messerly, 2014/02/04 23:40:41]

yeah, thanks for the idea :). It made the code a lot nicer in template_iterator
and down below in _listCombinator, which I had trouble understanding before!

+   * Gets the [i]th text entry. Note that [length] can be passed to get the
+   * final text entry.
+   */
+  String getText(int i) => _tokens[i * _TOKEN_SIZE + _TOKEN_TEXT];
+
+  /** Gets the oneTime flag for the [i]th token. */
+  bool getOneTime(int i) => _tokens[i * _TOKEN_SIZE + _TOKEN_ONETIME];
+
+  /** Gets the path for the [i]th token. */
+  PropertyPath getPath(int i) => _tokens[i * _TOKEN_SIZE + _TOKEN_PATH];
+
+  /** Gets the prepareBinding function for the [i]th token. */
+  Function getPrepareBinding(int i) =>
+      _tokens[i * _TOKEN_SIZE + _TOKEN_PREPAREFN];
+
+
+  /**
+   * Parses {{ mustache }} bindings.
+   *
+   * Returns null if there are no matches. Otherwise returns the parsed tokens.
+   */
+  static MustacheTokens parse(String s, String name, node,
+      BindingDelegate delegate) {
+    if (s == null || s.isEmpty) return null;
+
+    var tokens = null;
+    var length = s.length;
+    var lastIndex = 0;
+    var onlyOneTime = true;
+    while (lastIndex < length) {
+      var startIndex = s.indexOf('{{', lastIndex);
+      var oneTimeStart = s.indexOf('[[', lastIndex);
+      var oneTime = false;
+      var terminator = '}}';
+
+      if (oneTimeStart >= 0 &&
+          (startIndex < 0 || oneTimeStart < startIndex)) {
+        startIndex = oneTimeStart;
+        oneTime = true;
+        terminator = ']]';
+      }
+
+      var endIndex = startIndex < 0 ? -1 : s.indexOf(terminator, startIndex + 2);

[justinfagnani, 2014/02/04 23:34:49]

long line

[Jennifer Messerly, 2014/02/04 23:40:41]

Done.

+
+      if (endIndex < 0) {
+        if (tokens == null) return null;
+
+        tokens.add(s.substring(lastIndex)); // TEXT
+        break;
+      }
+
+      if (tokens == null) tokens = [];
+      tokens.add(s.substring(lastIndex, startIndex)); // TEXT
+      var pathString = s.substring(startIndex + 2, endIndex).trim();
+      tokens.add(oneTime); // ONETIME?
+      onlyOneTime = onlyOneTime && oneTime;
+      tokens.add(new PropertyPath(pathString)); // PATH
+      var delegateFn = delegate == null ? null :
+          delegate.prepareBinding(pathString, name, node);
+      tokens.add(delegateFn);
+
+      lastIndex = endIndex + 2;
+    }
+
+    if (lastIndex == length) tokens.add('');
+
+    return new MustacheTokens._(tokens, onlyOneTime);
+  }
+
+
+  // Dart note: split "combinator" into the single/list variants, so the
+  // argument can be typed.
+  String _singleCombinator(Object value) {
+    if (value == null) value = '';
+    return '${getText(0)}$value${getText(length)}';
+  }
+
+  String _listCombinator(List<Object> values) {
+    var newValue = new StringBuffer(getText(0));
+    int len = this.length;
+    for (var i = 0; i < len; i++) {
+      var value = values[i];
+      if (value != null) newValue.write(value);
+      newValue.write(getText(i + 1));
+    }
+    return newValue.toString();
+  }
+}