Testsuite wobbles

[Simon Peyton Jones]

Friends
I'm seeing test-suite wobbles like these:

Unexpected failures:

   /tmp/ghctest-9ez8i8lx/test   spaces/./ghci.debugger/scripts/break006.run  break006 [bad stdout] (ghci)

   /tmp/ghctest-9ez8i8lx/test   spaces/./ghci.debugger/scripts/hist001.run   hist001 [bad stdout] (ghci)

   /tmp/ghctest-9ez8i8lx/test   spaces/./ghci.debugger/scripts/hist002.run   hist002 [bad stdout] (ghci)

   /tmp/ghctest-9ez8i8lx/test   spaces/./profiling/should_run/scc001.run     scc001 [bad profile] (profasm)
It turns out that they are all variations in the order of output, e.g.

--- ./scripts/break006.run/break006.stdout.normalised 2018-05-01 13:20:13.297808002 +0100

+++ ./scripts/break006.run/break006.run.stdout.normalised   2018-05-01 13:20:13.297808002 +0100

@@ -4,14 +4,14 @@

x :: Integer = 1

xs :: [Integer] = [2,3]

xs :: [Integer] = [2,3]

-f :: Integer -> a = _

x :: Integer = 1

+f :: Integer -> a = _

_result :: [a] = _

y = (_t1::a)

y = 2

xs :: [Integer] = [2,3]

-f :: Integer -> Integer = _

x :: Integer = 1

+f :: Integer -> Integer = _

_result :: [Integer] = _

y :: Integer = 2

_t1 :: Integer = 2

*** unexpected failure for break006(ghci)

I've seen this go to and fro recently.    It looks as if the output is not deterministic somehow.  (The order in which the bindings are printed is immaterial.)
Does this happen for others?  Does the master CI build show this problem?
I don't know whether to commit these changes, because I don't know if it affects anyone else.  And it's certainly tiresome for me.
Any ideas?
Simon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-devs/attachments/20180501/310e695e/attachment.html>