Do not escape the ampersand "&" in entity references (bug #568485).

* codegen/docextract_to_xml.py (escape_text): Do not escape the
ampersand "&" in entity references (bug #568485).  Replace some
unusual entity references in the output with their literal values.

svn path=/trunk/; revision=998

2 files changed, 18 insertions, 7 deletions

diff --git a/ChangeLog b/ChangeLog
index 87fc720..c3f6f16 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,9 @@
+2009-01-27  Daniel Elstner  <danielk@openismus.com>
+
+	* codegen/docextract_to_xml.py (escape_text): Do not escape the
+	ampersand "&" in entity references (bug #568485).  Replace some
+	unusual entity references in the output with their literal values.
+
 2009-01-14  Paul Pogonyshev  <pogonyshev@gmx.net>
 
 	Bug 566571 – gtk.Buildable interface method override is not
diff --git a/codegen/docextract_to_xml.py b/codegen/docextract_to_xml.py
index 76ac85d..f8d3bae 100755
--- a/codegen/docextract_to_xml.py
+++ b/codegen/docextract_to_xml.py
@@ -7,21 +7,26 @@
 # # ./docextract_to_xml.py -s /gnome/head/cvs/gtk+/gtk/ -s /gnome/head/cvs/gtk+/docs/reference/gtk/tmpl/ > gtk_docs.xml
 
 import getopt
+import re
 import string
 import sys
 
 import docextract
 
 def escape_text(unescaped_text):
-    escaped_text = unescaped_text
+    # Escape every "&" not part of an entity reference
+    escaped_text = re.sub(r'&(?![A-Za-z]+;)', '&amp;', unescaped_text)
+
+    # These weird entities turn up in the output...
+    escaped_text = string.replace(escaped_text, '&mdash;', '&#8212;')
+    escaped_text = string.replace(escaped_text, '&ast;', '*')
+    escaped_text = string.replace(escaped_text, '&percnt;', '%')
+    escaped_text = string.replace(escaped_text, '&commat;', '@')
+
+    # Escape for both tag contents and attribute values
     escaped_text = string.replace(escaped_text, '<', '&lt;')
     escaped_text = string.replace(escaped_text, '>', '&gt;')
-    escaped_text = string.replace(escaped_text, '&', '&amp;')
-    escaped_text = string.replace(escaped_text, '\'', '&apos;')
-    escaped_text = string.replace(escaped_text, '\"', '&quot;')
-
-    #Apparently this is an undefined symbol:
-    escaped_text = string.replace(escaped_text, '&mdash;', ' mdash ')
+    escaped_text = string.replace(escaped_text, '"', '&quot;')
 
     return escaped_text