TSRM should set a smarter value for expected_threads #10854
Description
Description
The tsrm_startup() function is currently always called with expected_threads = 1. This means that the hashtable used in the TSRM will only contain a single bucket, and all thread resources will therefore be in the same linked list. So it's not really a hashtable atm, even though it's supposed to be.
We could improve performance in principle by choosing a more appropriate value for expected_threads, by somehow letting the SAPI give an estimation based on its configuration.
I noticed this while working on #10737.