Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Fixing issue with Windows timeout test (#1946)
Summary: Context and explanation: Timing works differently on Windows, and its imprecision has caused test failures in the past. When we started testing against Python 3.11, Windows tests involving timing started failing flakily. Adding some print statements showed that when the tests failed, time elapsed was measured as 0.0 seconds when it should have been substantially more. This is odd, since timing (including `monotonic`) was supposed to be more precise in Python 3.11. I couldn't debug this too thoroughly since I don't have access to a Windows machine outside of GitHub Actions. The fix: Increased `sleep` from 1e-3s to 1e-2s in a test where we are making sure something times out within a limit of 1e-4s. Pull Request resolved: #1946 Test Plan: I'm going to run units a few times to ensure that they pass consistently. Reviewed By: saitcakmak Differential Revision: D47810397 Pulled By: esantorella fbshipit-source-id: 323375904553c042b84ea480f3e9eb50e6b5e21e
- Loading branch information