How to execute jobs in parallel potential Stackoverflow ?

Hi @dbader ,

First of all thank you for creating this also package, it is very useful and versatile.

I have been using the following feature for little over a year now,until recently I have noticed some issue:
https://schedule.readthedocs.io/en/stable/faq.html#how-to-execute-jobs-in-parallel

Recently, the number of thread that I'm running has been increased(it was about 10-20 pending thread, now it about 40+ pending thread), then the `schedule.run_pending()` will not kick off any task afterwards, more strangely, there is no error or exception has been through, it felt that the whole method call been silenced/skipped.

I did some initial investigation, noticed the thread limit might be the issue, not the number of thread but the `stack size`. I'm running it on MacOS, the default stack size is about 8Mb, the max size the OS allowed is 65532K, which is what I set for now, after change the Stack Size to the Max. the issue seems went away at the moment.

So, my question goes, is there a potential stack overflow issue while doing multi-threading with many task running in parallel ? If is, why it didn't blow up... with limited knowledge of system programming, I'm hoping someone know about it can help answer it.

System: Mac OS 10.14.3
Python: 3.7.2
Schedule Version: 0.6.0

Start with the documented parallel-execution approach in the FAQ and the schedule.run_pending() entry point. Reproduce the behavior on macOS 10.14.3 with Python 3.7.2, Schedule 0.6.0, and 40+ pending threads while comparing stack-size settings; done means determining whether the library is responsible and documenting or resolving the observed silent stop.

Written by the indexing model from the issue text.