4 ms·
almost like people should test their code before using a new python version in production.
by danbala 4y ago
almost like people should test their code before using a new python version in production.
- klibertp 4y agoThat's the problem: the kind of issues the GIL prevents is notoriously hard to reproduce and debug. It would probably pass 90% of the test suites out there, unless you wrote yours with multi-threading in mind. Same for empirical tests by QA teams - they rarely cause enough load on the system for the bugs to surface. I might be too pessimistic here, but this really looks to me like a disaster waiting to happen...
- gcbirzan 4y agoBecause concurrency issues are notoriously easy to catch.
- alkonaut 4y agoTesting can’t prove the absence of bugs.