[Lxml-checkins] r45025 - lxml/trunk/src/lxml

[scoder at codespeak.net]

Author: scoder
Date: Fri Jul 13 15:42:56 2007
New Revision: 45025

Modified:
   lxml/trunk/src/lxml/parser.pxi
Log:
work around libxml2 not being able to detect BOM-less UTF-16LE

Modified: lxml/trunk/src/lxml/parser.pxi
==============================================================================
--- lxml/trunk/src/lxml/parser.pxi	(original)
+++ lxml/trunk/src/lxml/parser.pxi	Fri Jul 13 15:42:56 2007
@@ -156,6 +156,15 @@
     l = python.PyUnicode_GET_DATA_SIZE(utext)
     buffer = python.PyUnicode_AS_DATA(utext)
     enc = _findEncodingName(buffer, l)
+    if enc == NULL:
+        # apparently, libxml2 can't detect UTF16LE on some systems
+        if l >= 4 and \
+               buffer[0] == c'<' and buffer[1] == c'\0' and \
+               buffer[2] == c't' and buffer[3] == c'\0':
+            enc = "UTF16LE"
+        else:
+            # not my fault, it's YOUR broken system :)
+            return
     enchandler = tree.xmlFindCharEncodingHandler(enc)
     if enchandler is not NULL:
         global _UNICODE_ENCODING
@@ -174,6 +183,8 @@
         return "UCS-4LE"
     elif enc == tree.XML_CHAR_ENCODING_UCS4BE:
         return "UCS-4BE"
+    elif enc == tree.XML_CHAR_ENCODING_NONE:
+        return NULL
     else:
         return tree.xmlGetCharEncodingName(enc)