Open mlao-pdx opened 5 years ago
This is unfortunate.. But interesting! :) Nice find.
This specific piece of the fuzzing engine is being overhauled in the next version of BT so things should straighten themselves out soon.
Seeing this helps a lot though and really helps nail in a couple decisions for the next version.
Thanks for investigating!
You are welcome. Hope it helps to make a better product for all of us.
Thank you for creating battle-tested.
Hi,
Not sure what is causing the issue, but trying to run the below on a Mac with Python 3.7 virtual env with Hydrogen/Atom, and battle-tested ==2017.11.18.1, I run into issues. The test count (here 1752) creeps up with each run with slightly different increments, probably due to the test case shrinking and caching that happens underneath the sheets with Hypothesis. When setting the count to or below the achieved test count, say 1700), it runs without error. Slightly above the previously achieved test count, say 1760, I might get lucky. Far above, like the 10K, then it will fail.
Arnoud.
Pipfile:
Code run and its redacted result: