On Wed, 31 May 2023 07:38:39 GMT, Aleksey Shipilev <sh...@openjdk.org> wrote:
>> test/jdk/java/util/UUID/UUIDTest.java line 79:
>>
>>> 77: }
>>> 78: if (!set.add(u)) {
>>> 79: throw new Exception("UUID collision: " + u);
>>
>> I would be concerned that if this failure was reported, it would be
>> intermittent, hard to track down, and not reproducible. Without a hook for
>> the generator and the seed, its just going to be noise in the testing.
>
> This is a non-practical concern, IMO. By spec, `UUID.randomUUID` is generated
> from the cryptographically secure random, with >120 bits of randomness, so
> the collision is extremely unlikely. Collision math involves birthday
> paradox, but Wikipedia article on UUIDs fortunately gives us the approximated
> solutions already:
> https://en.wikipedia.org/wiki/Universally_unique_identifier#Collisions
>
> Quote: "Thus, the probability to find a duplicate within 103 trillion
> version-4 UUIDs is one in a billion."
>
> In other words, finding a collision in this test with 1M UUIDs points to the
> implementation issue, not a test bug, with a very high probability. In yet
> another words, if a unit test with 1M UUIDs is able to find a collision, then
> this is a strong signal that many production systems that assume extremely
> low collision probability are up for subtle misbehavior.
My point was that its probably not practical to test (more than once).
If it fails, it will be considered just as you propose and disregarded and in
the meantime consumes test cycles in each of the test contexts. Either provide
more information about the conditions under which it failed or remove it.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/14134#discussion_r1211696030
