Handle unicode decoding of commit messages in bzr-fastimport, python-fastimport no longer takes care of this.

1 files changed, 6 insertions, 1 deletions

diff --git a/bzr_commit_handler.py b/bzr_commit_handler.py
index c47a39d..1df4c86 100644
--- a/bzr_commit_handler.py
+++ b/bzr_commit_handler.py
@@ -261,7 +261,12 @@ class GenericCommitHandler(processor.CommitHandler):
         self._save_author_info(rev_props)
         committer = self.command.committer
         who = self._format_name_email(committer[0], committer[1])
-        message = self.command.message
+        try:
+            message = self.command.message.decode("utf-8")
+        except UnicodeDecodeError:
+            self.warning(
+                "commit message not in utf8 - replacing unknown characters")
+            message = message.decode('utf-8', 'replace')
         if not _serializer_handles_escaping:
             # We need to assume the bad ol' days
             message = helpers.escape_commit_message(message)