id,summary,reporter,owner,description,type,status,priority,component,severity,resolution,keywords,cc
129,Incorrect Unicode BOM generation,Valentin Nechayev <netchv@…>,xi,"py-YAML 3.07, with Python 2.5 and FreeBSD (package name py25-yaml-3.07_2)

When yaml.dump() generates stream in utf-16be or utf-16le, it generates byte-order mark (BOM), but makes it incorrectly. Example:

{{{
>>> yaml.dump(""xyz"", encoding = 'utf-16be')
'\x00\xff\x00\xfe\x00x\x00y\x00z\x00\n\x00.\x00.\x00.\x00\n'
}}}

Instead, it should generate:

{{{
'\xfe\xff\x00x\x00y\x00z\x00\n\x00.\x00.\x00.\x00\n'
}}}

Fix:

{{{
--- 01/PyYAML-3.07/lib/yaml/emitter.py  2008-12-29 01:36:32.000000000 +0200
+++ work/PyYAML-3.07/lib/yaml/emitter.py        2009-06-06 16:48:39.000000000 +0300
@@ -787,7 +787,7 @@
     def write_stream_start(self):
         # Write BOM if needed.
         if self.encoding and self.encoding.startswith('utf-16'):
-            self.stream.write(u'\xFF\xFE'.encode(self.encoding))
+            self.stream.write(u'\uFEFF'.encode(self.encoding))
 
     def write_stream_end(self):
         self.flush_stream()
}}}

P.S. I guess it also should generate BOMs for utf-32*",defect,closed,normal,pyyaml,normal,fixed,,