Respond to review:

- Refactor both the base64 encoding and decoding into utility functions.

Also:

- Mechanically fix some other broken untested code.

5 files changed, 24 insertions, 24 deletions

diff --git a/cloudinit/config/cc_seed_random.py b/cloudinit/config/cc_seed_random.py
index 981e1b08..bb64b0f5 100644
--- a/cloudinit/config/cc_seed_random.py
+++ b/cloudinit/config/cc_seed_random.py
@@ -38,13 +38,7 @@ def _decode(data, encoding=None):
     if not encoding or encoding.lower() in ['raw']:
         return data
     elif encoding.lower() in ['base64', 'b64']:
-        # Try to give us a native string in both Python 2 and 3, and remember
-        # that b64decode() returns bytes in Python 3.
-        decoded = base64.b64decode(data)
-        try:
-            return decoded.decode('utf-8')
-        except UnicodeDecodeError:
-            return decoded
+        return util.b64d(data)
     elif encoding.lower() in ['gzip', 'gz']:
         return util.decomp_gzip(data, quiet=False)
     else:
diff --git a/cloudinit/config/cc_ssh_authkey_fingerprints.py b/cloudinit/config/cc_ssh_authkey_fingerprints.py
index 51580633..6ce831bc 100644
--- a/cloudinit/config/cc_ssh_authkey_fingerprints.py
+++ b/cloudinit/config/cc_ssh_authkey_fingerprints.py
@@ -32,7 +32,7 @@ from cloudinit import util
 
 def _split_hash(bin_hash):
     split_up = []
-    for i in xrange(0, len(bin_hash), 2):
+    for i in range(0, len(bin_hash), 2):
         split_up.append(bin_hash[i:i + 2])
     return split_up
 
diff --git a/cloudinit/sources/DataSourceOpenNebula.py b/cloudinit/sources/DataSourceOpenNebula.py
index a0275cda..61709c1b 100644
--- a/cloudinit/sources/DataSourceOpenNebula.py
+++ b/cloudinit/sources/DataSourceOpenNebula.py
@@ -426,12 +426,7 @@ def read_context_disk_dir(source_dir, asuser=None):
                                context.get('USER_DATA_ENCODING'))
         if encoding == "base64":
             try:
-                userdata = base64.b64decode(results['userdata'])
-                # In Python 3 we still expect a str, but b64decode will return
-                # bytes.  Convert to str.
-                if isinstance(userdata, bytes):
-                    userdata = userdata.decode('utf-8')
-                results['userdata'] = userdata
+                results['userdata'] = util.b64d(results['userdata'])
             except TypeError:
                 LOG.warn("Failed base64 decoding of userdata")
 
diff --git a/cloudinit/sources/DataSourceSmartOS.py b/cloudinit/sources/DataSourceSmartOS.py
index f59ad3d6..9d48beab 100644
--- a/cloudinit/sources/DataSourceSmartOS.py
+++ b/cloudinit/sources/DataSourceSmartOS.py
@@ -351,16 +351,7 @@ def query_data(noun, seed_device, seed_timeout, strip=False, default=None,
 
     if b64:
         try:
-            # Generally, we want native strings in the values.  Python 3's
-            # b64decode will return bytes though, so decode them to utf-8 if
-            # possible.  If that fails, return the bytes.
-            decoded = base64.b64decode(resp)
-            try:
-                if isinstance(decoded, bytes):
-                    return decoded.decode('utf-8')
-            except UnicodeDecodeError:
-                pass
-            return decoded
+            return util.b64d(resp)
         # Bogus input produces different errors in Python 2 and 3; catch both.
         except (TypeError, binascii.Error):
             LOG.warn("Failed base64 decoding key '%s'", noun)
diff --git a/cloudinit/util.py b/cloudinit/util.py
index 766f8e32..8916cc11 100644
--- a/cloudinit/util.py
+++ b/cloudinit/util.py
@@ -44,6 +44,7 @@ import sys
 import tempfile
 import time
 
+from base64 import b64decode, b64encode
 from six.moves.urllib import parse as urlparse
 
 import six
@@ -90,6 +91,25 @@ def encode_text(text, encoding='utf-8'):
         return text
     return text.encode(encoding)
 
+
+def b64d(source):
+    # Base64 decode some data, accepting bytes or unicode/str, and returning
+    # str/unicode if the result is utf-8 compatible, otherwise returning bytes.
+    decoded = b64decode(source)
+    if isinstance(decoded, bytes):
+        try:
+            return decoded.decode('utf-8')
+        except UnicodeDecodeError:
+            return decoded
+
+def b64e(source):
+    # Base64 encode some data, accepting bytes or unicode/str, and returning
+    # str/unicode if the result is utf-8 compatible, otherwise returning bytes.
+    if not isinstance(source, bytes):
+        source = source.encode('utf-8')
+    return b64encode(source).decode('utf-8')
+
+
 # Path for DMI Data
 DMI_SYS_PATH = "/sys/class/dmi/id"