Clean up histograms.xml formatting

tools/metrics/common/pretty_print_xml.py

 # Copyright 2014 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 """Utility file for pretty printing xml file.
 
 The function PrettyPrintXml will be used for formatting both histograms.xml
 and actions.xml.
 """
 
@@ skipping 48 matching lines @@
       paragraphs[-1].append(l)
   # Remove trailing empty paragraph if present.
   if paragraphs and not paragraphs[-1]:
     paragraphs = paragraphs[:-1]
   return ['\n'.join(p) for p in paragraphs]
 
 
 class XmlStyle(object):
   """A class that stores all style specification for an output xml file."""
 
-  def __init__(self, attribute_order, tags_that_have_extra_newline,
-               tags_that_dont_indent, tags_that_allow_single_line,
-               tags_alphabetization_rules):
+  def __init__(self, attribute_order, required_attributes,
+               tags_that_have_extra_newline, tags_that_dont_indent,
+               tags_that_allow_single_line, tags_alphabetization_rules):
     self.attribute_order = attribute_order
+    self.required_attributes = required_attributes
     self.tags_that_have_extra_newline = tags_that_have_extra_newline
     self.tags_that_dont_indent = tags_that_dont_indent
     self.tags_that_allow_single_line = tags_that_allow_single_line
     self.tags_alphabetization_rules = tags_alphabetization_rules
 
   def PrettyPrintXml(self, tree):
     tree = self._TransformByAlphabetizing(tree)
     tree = self.PrettyPrintNode(tree)
     return tree
 
@@ skipping 109 matching lines @@
       newlines_after_open, newlines_before_close, newlines_after_close = (
           self.tags_that_have_extra_newline.get(node.tagName, (1, 1, 0)))
       # Open the tag.
       s = ' ' * indent + '<' + node.tagName
 
       # Calculate how much space to allow for the '>' or '/>'.
       closing_chars = 1
       if not node.childNodes:
         closing_chars = 2
 
+      attributes = node.attributes.keys()
+      required_attributes = [attribute for attribute in self.required_attributes
+                             if attribute in self.attribute_order[node.tagName]]
+      missing_attributes = [attribute for attribute in required_attributes
+                            if attribute not in attributes]
+
+      for attribute in missing_attributes:
+        logging.error(
+            'Missing attribute "%s" in tag "%s"', attribute, node.tagName)
+      if missing_attributes:
+        raise Error(
+            'Missing attributes {0} in tag "{1}"'.format(
+                ', '.join('"{0}"'.format(a) for a in missing_attributes),

[Alexei Svitkine (slow), 2017/05/19 20:05:28]

Nit: I'd extract this inner statement into a var above so that this statement is
a bit easier to parse.

+                node.tagName))
+
       # Pretty-print the attributes.
-      attributes = node.attributes.keys()
       if attributes:
         # Reorder the attributes.
         unrecognized_attributes = (
             [a for a in attributes
              if a not in self.attribute_order[node.tagName]])
         attributes = [a for a in self.attribute_order[node.tagName]
                       if a in attributes]
 
         for a in unrecognized_attributes:
           logging.error(
@@ skipping 63 matching lines @@
 
     # Handle comment nodes.
     if node.nodeType == xml.dom.minidom.Node.COMMENT_NODE:
       return '<!--%s-->\n' % node.data
 
     # Ignore other node types. This could be a processing instruction
     # (<? ... ?>) or cdata section (<![CDATA[...]]!>), neither of which are
     # legal in the histograms XML at present.
     logging.error('Ignoring unrecognized node data: %s', node.toxml())
     raise Error('Ignoring unrecognized node data: {0}'.format(node.toxml()))