I'm seeing some unrelated test failures on PPC Linux (big endian) in my PEP-456 branch.

http://buildbot.python.org/all/builders/PPC64%20PowerLinux%20custom/builds/17/steps/test/logs/stdio

======================================================================
FAIL: test_frozensets (test.test_gdb.PrettyPrintTests)
Verify the pretty-printing of frozensets
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/shager/cpython-buildarea/custom.edelsohn-powerlinux-ppc64/build/Lib/test/test_gdb.py", line 331, in test_frozensets
    self.assertGdbRepr(frozenset(['a', 'b']), "frozenset({'a', 'b'})")
  File "/home/shager/cpython-buildarea/custom.edelsohn-powerlinux-ppc64/build/Lib/test/test_gdb.py", line 231, in assertGdbRepr
    % (gdb_repr, exp_repr, gdb_output)))
AssertionError: "frozenset({'b', 'a'})" != "frozenset({'a', 'b'})"
- frozenset({'b', 'a'})
?             ^    ^
+ frozenset({'a', 'b'})
?             ^    ^


======================================================================
FAIL: test_sets (test.test_gdb.PrettyPrintTests)
Verify the pretty-printing of sets
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/shager/cpython-buildarea/custom.edelsohn-powerlinux-ppc64/build/Lib/test/test_gdb.py", line 316, in test_sets
    self.assertGdbRepr(set(['a', 'b']), "{'a', 'b'}")
  File "/home/shager/cpython-buildarea/custom.edelsohn-powerlinux-ppc64/build/Lib/test/test_gdb.py", line 231, in assertGdbRepr
    % (gdb_repr, exp_repr, gdb_output)))
AssertionError: "{'b', 'a'}" != "{'a', 'b'}"
- {'b', 'a'}
+ {'a', 'b'}

Also on System Z:

http://buildbot.python.org/all/builders/System%20Z%20Linux%203.x/builds/1009/steps/test/logs/stdio

Setting priority to "normal", since it's the only test failing on
System Z and generally green buildbots are more useful.

Oh, I didn't notice this issue. I created the duplicate #19753 and I did some changes to try to fix it.

When trying to workaround a bug in the implementation of the PEP 456 (which was not known as a bug at this time), I implemented something using subprocessing which may be reused on SystemZ (or maybe on all platforms):

changeset:   87290:11cb1c8faf11
user:        Victor Stinner <victor.stinner@gmail.com>
date:        Wed Nov 20 12:27:48 2013 +0100
files:       Lib/test/test_gdb.py
description:
Issue #19183: Fix repr() tests of test_gdb, hash() is now platform dependent

> http://buildbot.python.org/all/builders/System%20Z%20Linux%203.x/builds/1009/steps/test/logs/stdio

This is a different issue than the one reported here.

I'm not seeing any recent failures on "PPC64 PowerLinux 3.x". Christian, can we close?