Open linearray opened 1 year ago
Another one: GHC 9.4.4 on macOS on amd64:
...Internal.Stream:
xor munges input: [Failed]
*** Failed! Falsified (after 1 test):
Message "\NUL"
(used seed -4842425483227000893)
same thing on arm
...Internal.Stream:
Stream is apropriately sized: [Arguments exhausted after 16885 tests]
xor munges input: [Failed]
*** Failed! Falsified (after 1 test):
Message "\NUL"
(used seed 4324022072132119234)
The failures in the xor
test were in all likelihood due to chance, because they all happened on a 1-byte bytestring a
, and it's not unlikely to randomly generate another byte b
such that a xor b = a
(that b
is 00000000
of course). Fixed in https://github.com/tel/saltine/commit/a0919e4d3e44b8e8e838d9120a8d85c24f17d7b2
For now that leaves the failed test from the beginning of the issue.
https://github.com/tel/saltine/actions/runs/4007439690/jobs/7280132376
With ghc 9.2.2:
We had a number of seemingly randomly failing tests in the past, maybe we can find out what is going on here.