random.gauss() uses a piece of hidden state used by nothing else,

and the .seed() and .whseed() methods failed to reset it.  In other
words, setting the seed didn't completely determine the sequence of
results produced by random.gauss().  It does now.  Programs repeatedly
mixing calls to a seed method with calls to gauss() may see different
results now.

Bugfix candidate (random.gauss() has always been broken in this way),
despite that it may change results.

Lib/random.py

@@ -116,7 +116,6 @@ class Random:
         """
 
         self.seed(x)
-        self.gauss_next = None
 
 ## -------------------- core generator -------------------
 
@@ -150,6 +149,8 @@ class Random:
         a, z = divmod(a, 30322)
         self._seed = int(x)+1, int(y)+1, int(z)+1
 
+        self.gauss_next = None
+
     def random(self):
         """Get the next random number in the range [0.0, 1.0)."""
 
@@ -238,6 +239,8 @@ class Random:
         # Zero is a poor seed, so substitute 1
         self._seed = (x or 1, y or 1, z or 1)
 
+        self.gauss_next = None
+
     def whseed(self, a=None):
         """Seed from hashable object's hash code.
 

Lib/test/test_random.py

@@ -0,0 +1,19 @@
+import test_support
+import random
+
+# Ensure that the seed() method initializes all the hidden state.  In
+# particular, through 2.2.1 it failed to reset a piece of state used by
+# (and only by) the .gauss() method.
+
+for seed in 1, 12, 123, 1234, 12345, 123456, 654321:
+    for seeder in random.seed, random.whseed:
+        seeder(seed)
+        x1 = random.random()
+        y1 = random.gauss(0, 1)
+
+        seeder(seed)
+        x2 = random.random()
+        y2 = random.gauss(0, 1)
+
+        test_support.vereq(x1, x2)
+        test_support.vereq(y1, y2)

Misc/NEWS

@@ -104,6 +104,13 @@ Extension modules
 
 Library
 
+- random.gauss() uses a piece of hidden state used by nothing else,
+  and the .seed() and .whseed() methods failed to reset it.  In other
+  words, setting the seed didn't completely determine the sequence of
+  results produced by random.gauss().  It does now.  Programs repeatedly
+  mixing calls to a seed method with calls to gauss() may see different
+  results now.
+
 - The pickle.Pickler class grew a clear_memo() method to mimic that
   provided by cPickle.Pickler.