Closed non closed 9 months ago
Thanks for this! Good catch
@non Have a look at the above PR -- let me know how this works for you. Thanks again!
@non If you have a moment, check out the PR mentioned above, and close the issue if it's resolved for you. Thanks!
Closing as fix should be on HEAD. Re-open if it doesn't work for you.
Here's an example test that demonstrates the problem:
On my machine this erroneous test always passes (the value generated for
f(0)
is always58927336
). I would expect half of the generated functions to produce an odd output forf(0)
.The problem is that when generating a function, the seed is ignored. The lambda generator hashes its inputs to produce outputs, which means all generated functions work the same way. The easiest fix would be to combine the original generator seed (currently unused) with the hashed arguments to produce an output. This would ensure that two generated function would likely behave differently for the same input.