Creates thousands of threads until it crashes on Linux

[MortenLohne]

Calling Scope::zoom() or Scope::recurse() from inside a job seems to create new system threads on Linux. I don't really know what's going on, does every waiting job gets its own thread? Anyway, thousands of threads are created until eventually the OS gives up and it crashes with "thread '' panicked at 'called Result::unwrap() on an Err value: Error { repr: Os { code: 11, message: "Resource temporarily unavailable" } }"
Relevant stacktrace:

   8:     0x557f33a19e1d - core::panicking::panic_fmt::he9c7f335d160b59d
                               at /checkout/src/libcore/panicking.rs:69
   9:     0x557f33995dfd - core::result::unwrap_failed::hf35eb5c4700747f4
  10:     0x557f3398af6f - scoped_pool::Pool::expand::h7f75d890a03c4eca
  11:     0x557f3398aa0f - scoped_pool::Pool::with_thread_config::hdf014a2bf1905926
  12:     0x557f33990ad7 - std::panicking::try::do_call::h87060de474abd462

[reem]

Judging by that stacktrace, the actual panic occurs in Pool::with_thread_config - are you calling that function? If so with what arguments?

This is pretty weird behavior, if you have a reproduction of some kind I would really appreciate it.

[dearsxx0918]

I also found this issue, the scoped pool use too much memory, seems it will not release the threads its created.