Running into this issue on an amd threadripper system: https://bugs.python.org/issue26903
This manifests as 'ValueError: need at most 63 handles, got a sequence of length 65'
It looks like this can be solved on my machine by setting n_processes=60 instead of n_processes=none (which then defaults to 64, causing the crash), e.g. in demo_initialize_ca.py. Since this seems to be caused by an external unfixed bug, might it make sense to default the N here to 60 generically?
Running into this issue on an amd threadripper system: https://bugs.python.org/issue26903
This manifests as 'ValueError: need at most 63 handles, got a sequence of length 65'
It looks like this can be solved on my machine by setting n_processes=60 instead of n_processes=none (which then defaults to 64, causing the crash), e.g. in demo_initialize_ca.py. Since this seems to be caused by an external unfixed bug, might it make sense to default the N here to 60 generically?