Start generating translated policy templates

tools/grit/grit/format/policy_templates/template_formatter.py

Index: tools/grit/grit/format/policy_templates/template_formatter.py
diff --git a/tools/grit/grit/format/policy_templates/template_formatter.py b/tools/grit/grit/format/policy_templates/template_formatter.py
index 0e304afd1d6d5b3e59551cf80b6cdf459e4fef29..ad1ebe8f53b2ff3b82ca8c9e04e3a70034de8863 100644
--- a/tools/grit/grit/format/policy_templates/template_formatter.py
+++ b/tools/grit/grit/format/policy_templates/template_formatter.py
@@ -94,7 +94,10 @@ class TemplateFormatter(interface.ItemFormatter):
     if (isinstance(item, structure.StructureNode) and
         item.attrs['type'] == 'policy_template_metafile'):
       assert self._policy_data == None
-      json_text = item.gatherer.Translate(self._lang)
+      json_text = item.gatherer.Translate(
+          self._lang,
+          pseudo_if_not_available=False,

[Jói, 2011/05/23 13:13:51]

pseudo_if_not_available and fallback_to_english are usually fetched via
item.PseudoAllowed() and item.ShouldFallbackToEnglish(), where they are
retrieved as attributes from the appropriate enclosing nodes
(fallback_to_english from the enclosing <structures>, <includes>, <messages> or
<identifiers> node, allow_pseudo from the <release> node).  Is there a reason
not to do that here?

[gfeher, 2011/05/23 20:55:10]

Done!

+          fallback_to_english=True)
       self._policy_data = eval(json_text)
     for child in item.children:
       self._ParseGritNodes(child)