gp_inf: Read/write files with a UTF-16LE BOM in GptTmpl.inf

Regression caused by 16596842a62bec0a9d974c48d64000e3c079254e

[MS-GPSB] 2.2 Message Syntax says that you have to write a BOM which I
didn't do up until this patch. UTF-16 as input encoding was marked much
higher up in the inheritance tree, which got overriden with the Python 3
fixes. I've now marked the encoding much more obviously for this file.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=14004

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>

Autobuild-User(master): Gary Lockyer <gary@samba.org>
Autobuild-Date(master): Fri Jul 19 02:20:47 UTC 2019 on sn-devel-184

python/samba/gp_parse/gp_inf.py

@@ -29,11 +29,11 @@ from samba.gp_parse import GPParser
 # [MS-GPSB] Security Protocol Extension
 class GptTmplInfParser(GPParser):
     sections = None
-    encoding = 'utf-16le'
+    encoding = 'utf-16'
+    output_encoding = 'utf-16le'
 
     class AbstractParam:
         __metaclass__ = ABCMeta
-        encoding = 'utf-16le'
 
         def __init__(self):
             self.param_list = []
@@ -333,7 +333,10 @@ class GptTmplInfParser(GPParser):
 
     def write_binary(self, filename):
         with codecs.open(filename, 'wb+',
-                         self.encoding) as f:
+                         self.output_encoding) as f:
+            # Write the byte-order mark
+            f.write(u'\ufeff')
+
             for s in self.sections:
                 self.sections[s].write_section(s, f)