X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/ee434dea5db0b02fe762ff80b2270cf268e84378..b352c3862814fe0bdd2b5a40b1dc8171474dbb48:/sdk/python/arvados/_normalize_stream.py

diff --git a/sdk/python/arvados/_normalize_stream.py b/sdk/python/arvados/_normalize_stream.py
index babcabc1a7..c72b82be1c 100644
--- a/sdk/python/arvados/_normalize_stream.py
+++ b/sdk/python/arvados/_normalize_stream.py
@@ -1,6 +1,15 @@
+# Copyright (C) The Arvados Authors. All rights reserved.
+#
+# SPDX-License-Identifier: Apache-2.0
+
 from __future__ import absolute_import
 from . import config
 
+import re
+
+def escape(path):
+    return re.sub(r'[\\:\000-\040]', lambda m: "\\%03o" % ord(m.group(0)), path)
+
 def normalize_stream(stream_name, stream):
     """Take manifest stream and return a list of tokens in normalized format.
 
@@ -12,7 +21,7 @@ def normalize_stream(stream_name, stream):
 
     """
 
-    stream_name = stream_name.replace(' ', '\\040')
+    stream_name = escape(stream_name)
     stream_tokens = [stream_name]
     sortedfiles = list(stream.keys())
     sortedfiles.sort()
@@ -34,7 +43,7 @@ def normalize_stream(stream_name, stream):
     for streamfile in sortedfiles:
         # Add in file segments
         current_span = None
-        fout = streamfile.replace(' ', '\\040')
+        fout = escape(streamfile)
         for segment in stream[streamfile]:
             # Collapse adjacent segments
             streamoffset = blocks[segment.locator] + segment.segment_offset
@@ -44,13 +53,13 @@ def normalize_stream(stream_name, stream):
                 if streamoffset == current_span[1]:
                     current_span[1] += segment.segment_size
                 else:
-                    stream_tokens.append("{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
+                    stream_tokens.append(u"{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
                     current_span = [streamoffset, streamoffset + segment.segment_size]
 
         if current_span is not None:
-            stream_tokens.append("{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
+            stream_tokens.append(u"{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
 
         if not stream[streamfile]:
-            stream_tokens.append("0:0:{0}".format(fout))
+            stream_tokens.append(u"0:0:{0}".format(fout))
 
     return stream_tokens