From: Torsten Werner <twerner@debian.org>
Date: Fri, 4 Mar 2011 08:07:12 +0000 (+0100)
Subject: Yield utf-8 encoded strings from DBBinary.scan_contents().
X-Git-Url: https://git.decadent.org.uk/gitweb/?a=commitdiff_plain;h=3dc0b40477fad44d76168e9b1dd96a9cbbfbf660;p=dak.git

Yield utf-8 encoded strings from DBBinary.scan_contents().

Signed-off-by: Torsten Werner <twerner@debian.org>
---

diff --git a/daklib/dbconn.py b/daklib/dbconn.py
index 5bff4186..87db1a35 100755
--- a/daklib/dbconn.py
+++ b/daklib/dbconn.py
@@ -516,11 +516,13 @@ class DBBinary(ORMObject):
         tar = TarFile.open(fileobj = dpkg.stdout, mode = 'r|')
         for member in tar.getmembers():
             if member.isfile():
+                name = normpath(member.name)
+                # enforce proper utf-8 encoding
                 try:
-                    name = member.name.decode('utf-8')
+                    name.decode('utf-8')
                 except UnicodeDecodeError:
-                    name = member.name.decode('iso8859-1')
-                yield normpath(name)
+                    name = name.decode('iso8859-1').encode('utf-8')
+                yield name
         tar.close()
         dpkg.stdout.close()
         dpkg.wait()