[#2375] Extract messages in format for ckan trans tag

ckan/lib/extract.py

@@ -1,6 +1,7 @@
 import re
 from genshi.filters.i18n import extract as extract_genshi
 from jinja2.ext import babel_extract as extract_jinja2
+import lib.jinja_extensions
 
 jinja_extensions = '''
                     jinja2.ext.do, jinja2.ext.with_,
@@ -12,6 +13,25 @@
                     ckan.lib.jinja_extensions.UrlForExtension
                    '''
 
+def jinja2_cleaner(fileobj, *args, **kw):
+    # We want to format the messages correctly and intercepting here seems
+    # the best location
+    # add our custom tags
+    kw['options']['extensions'] = jinja_extensions
+
+    raw_extract = extract_jinja2(fileobj, *args, **kw)
+
+    for lineno, func, message, finder in raw_extract:
+
+        if isinstance(message, basestring):
+            message = lib.jinja_extensions.regularise_html(message)
+        elif message is not None:
+            message = (lib.jinja_extensions.regularise_html(message[0])
+                       ,lib.jinja_extensions.regularise_html(message[1]))
+
+        yield lineno, func, message, finder
+
+
 def extract_ckan(fileobj, *args, **kw):
     ''' Determine the type of file (Genshi or Jinja2) and then call the
     correct extractor function.
@@ -22,12 +42,10 @@ def extract_ckan(fileobj, *args, **kw):
     source = fileobj.read()
     if re.search('genshi\.edgewall\.org', source):
         # genshi
-        extractor_function = extract_genshi
+        output = extract_genshi(fileobj, *args, **kw)
     else:
         # jinja2
-        extractor_function = extract_jinja2
-        # add our custom tags
-        kw['options']['extensions'] = jinja_extensions
+        output = jinja2_cleaner(fileobj, *args, **kw)
     # we've eaten the file so we need to get back to the start
     fileobj.seek(0)
-    return extractor_function(fileobj, *args, **kw)
+    return output