Fix stringifying morphologies to be deterministic

Previously we were stringifying complex values (lists, dicts) by
using whatever Python produces, but different runs and versions of
Python may convert them differently, particularly dicts. We now do
it manually.

1 files changed, 18 insertions, 4 deletions

diff --git a/morphlib/cachekeycomputer.py b/morphlib/cachekeycomputer.py
index a1c2ebd6..6ae2b107 100644
--- a/morphlib/cachekeycomputer.py
+++ b/morphlib/cachekeycomputer.py
@@ -89,12 +89,26 @@ class CacheKeyComputer(object):
         if kind == 'chunk':
             keys['ref'] = artifact.source.sha1
         elif kind in ('system', 'stratum'):
-            checksum = hashlib.sha1()
             morphology = artifact.source.morphology
-            getkey = lambda s: [ord(c) for c in s]
-            for key in sorted(morphology.keys(), key=getkey):
-                checksum.update('%s=%s' % (key, morphology[key]))
+            le_dict = dict((k,morphology[k]) for k in morphology.keys())
+            checksum = hashlib.sha1(self._stringify(le_dict))
             keys['morphology-sha1'] = checksum.hexdigest()
 
         return keys
 
+    def _stringify(self, value):
+        if type(value) in [str, unicode, int]:
+            return str(value)
+        elif value is None:
+            return ''
+        elif type(value) is list:
+            return '[' + ','.join(self._stringify(x) for x in value) + ']'
+        elif type(value) is dict:
+            keys = value.keys()
+            keys.sort(key=lambda s: [ord(c) for c in s])
+            pairs = ['%s:%s' % (self._stringify(k), self._stringify(value[k]))
+                     for k in keys]
+            return '{' + ','.join(pairs) + '}'
+        else: # pragma: no cover
+            raise NotImplementedError(
+                        'type %s is not stringified' % type(value))