Roll Observatory packages and add a roll script

packages/html/test/support.dart

Index: packages/html/test/support.dart
diff --git a/packages/html/test/support.dart b/packages/html/test/support.dart
new file mode 100644
index 0000000000000000000000000000000000000000..1b4c0b49e50b57ba4216d55a0cc535ba5b2da7d2
--- /dev/null
+++ b/packages/html/test/support.dart
@@ -0,0 +1,164 @@
+/// Support code for the tests in this directory.
+library support;
+
+import 'dart:io';
+import 'dart:collection';
+import 'package:path/path.dart' as path;
+import 'package:html/src/treebuilder.dart';
+import 'package:html/dom.dart';
+import 'package:html/dom_parsing.dart';
+
+typedef TreeBuilder TreeBuilderFactory(bool namespaceHTMLElements);
+
+Map _treeTypes;
+Map<String, TreeBuilderFactory> get treeTypes {
+  if (_treeTypes == null) {
+    // TODO(jmesserly): add DOM here once it's implemented
+    _treeTypes = {"simpletree": (useNs) => new TreeBuilder(useNs)};
+  }
+  return _treeTypes;
+}
+
+final testDataDir = Platform.script.resolve('data').toFilePath();
+
+Iterable<String> getDataFiles(String subdirectory) {
+  var dir = new Directory(path.join(testDataDir, subdirectory));
+  return dir.listSync().where((f) => f is File).map((f) => f.path);
+}
+
+// TODO(jmesserly): make this class simpler. We could probably split on
+// "\n#" instead of newline and remove a lot of code.
+class TestData extends IterableBase<Map> {
+  final String _text;
+  final String newTestHeading;
+
+  TestData(String filename, [this.newTestHeading = "data"])
+      // Note: can't use readAsLinesSync here because it splits on \r
+      : _text = new File(filename).readAsStringSync();
+
+  // Note: in Python this was a generator, but since we can't do that in Dart,
+  // it's easier to convert it into an upfront computation.
+  Iterator<Map> get iterator => _getData().iterator;
+
+  List<Map> _getData() {
+    var data = <String, String>{};
+    var key = null;
+    var result = <Map>[];
+    var lines = _text.split('\n');
+    // Remove trailing newline to match Python
+    if (lines.last == '') {
+      lines.removeLast();
+    }
+    for (var line in lines) {
+      var heading = sectionHeading(line);
+      if (heading != null) {
+        if (data.length > 0 && heading == newTestHeading) {
+          // Remove trailing newline
+          data[key] = data[key].substring(0, data[key].length - 1);
+          result.add(normaliseOutput(data));
+          data = <String, String>{};
+        }
+        key = heading;
+        data[key] = "";
+      } else if (key != null) {
+        data[key] = '${data[key]}$line\n';
+      }
+    }
+
+    if (data.length > 0) {
+      result.add(normaliseOutput(data));
+    }
+    return result;
+  }
+
+  /// If the current heading is a test section heading return the heading,
+  /// otherwise return null.
+  static String sectionHeading(String line) {
+    return line.startsWith("#") ? line.substring(1).trim() : null;
+  }
+
+  static Map normaliseOutput(Map data) {
+    // Remove trailing newlines
+    data.forEach((key, value) {
+      if (value.endsWith("\n")) {
+        data[key] = value.substring(0, value.length - 1);
+      }
+    });
+    return data;
+  }
+}
+
+/// Serialize the [document] into the html5 test data format.
+testSerializer(document) {
+  return (new TestSerializer()..visit(document)).toString();
+}
+
+/// Serializes the DOM into test format. See [testSerializer].
+class TestSerializer extends TreeVisitor {
+  final StringBuffer _str;
+  int _indent = 0;
+  String _spaces = '';
+
+  TestSerializer() : _str = new StringBuffer();
+
+  String toString() => _str.toString();
+
+  int get indent => _indent;
+
+  set indent(int value) {
+    if (_indent == value) return;
+
+    var arr = new List<int>(value);
+    for (int i = 0; i < value; i++) {
+      arr[i] = 32;
+    }
+    _spaces = new String.fromCharCodes(arr);
+    _indent = value;
+  }
+
+  void _newline() {
+    if (_str.length > 0) _str.write('\n');
+    _str.write('|$_spaces');
+  }
+
+  visitNodeFallback(Node node) {
+    _newline();
+    _str.write(node);
+    visitChildren(node);
+  }
+
+  visitChildren(Node node) {
+    indent += 2;
+    for (var child in node.nodes) visit(child);
+    indent -= 2;
+  }
+
+  visitDocument(node) {
+    indent += 1;
+    for (var child in node.nodes) visit(child);
+    indent -= 1;
+  }
+
+  visitDocumentFragment(DocumentFragment node) => visitDocument(node);
+
+  visitElement(Element node) {
+    _newline();
+    _str.write(node);
+    if (node.attributes.length > 0) {
+      indent += 2;
+      var keys = new List.from(node.attributes.keys);
+      keys.sort((x, y) => x.compareTo(y));
+      for (var key in keys) {
+        var v = node.attributes[key];
+        if (key is AttributeName) {
+          AttributeName attr = key;
+          key = "${attr.prefix} ${attr.name}";
+        }
+        _newline();
+        _str.write('$key="$v"');
+      }
+      indent -= 2;
+    }
+    visitChildren(node);
+  }
+}