Kyle Stanley <aeros...@gmail.com> added the comment:
> Is the new asyncio.loop.shutdown_default_executor() suitable for event loops > that are run in single-step mode? I honestly can't say for certain; the primary intended use case for shutdown_default_executor() was to provide a means of properly finalizing the resources associated with the default executor without blocking the event loop, notably the threads in the ThreadPoolExecutor (which were intermittently left dangling). So, when working on the implementation w/ Yury and Andrew, I did not strongly consider single-step event loops. I was more concerned with how it fit in with asyncio.run() and safe finalization of executor resources. See https://bugs.python.org/issue34037 for context. If you have a recommendation for a change to the current version shutdown_default_executor() that would help provide compatibility with single-step event loops without hindering the primary goals, I'm sure it would be considered. > Also, what happens to pending executor futures? When using `loop.shutdown_default_executor()`, it calls executor.shutdown(wait=True), which waits for submitted futures to the executor to complete before joining the executor's workers (regardless of whether they're threads or processes). So, the executor should not be terminated prior to the pending futures being completed. >From a glance at the example code posted above, it seems like it would be >incompatible with asyncio.run(), which is a requirement for >shutdown_default_executor(). See >https://github.com/python/cpython/blob/b9c46a2c2d7fc68457bff641f78932d66f5e5f59/Lib/asyncio/runners.py#L8. ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue40357> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
