Replaces OSSpinLock with pthread_mutex_t

[akkrat]

Fixing #1785
pthread_mutex_t was chosen because

1. it's already using on some platforms
2. os_unfair_lock must be allocated on the heap so it couldn't be used as c++ class member (Concurrent Programming With GCD in Swift 3 or SO)

[dinhvh]

pthread_mutex_t is too slow.
We need to use something faster on iOS / macOS.

[akkrat]

libetpan uses pthread_mutex_t too
The other alternatives (such as dispatch semaphore or os_unfair_lock) are not much faster than pthread_mutex_t

[dinhvh]

https://gist.github.com/steipete/36350a8a60693d440954b95ea6cbbafc In my opnion os_unfair_lock would work.

…

On Mon, Sep 5, 2022 at 8:57 AM Artur Protska ***@***.***> wrote: libetpan uses pthread_mutex_t too The other alternatives (such as dispatch semaphore or os_unfair_lock) are not much faster than pthread_mutex_t — Reply to this email directly, view it on GitHub <#1953 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACRDCGCM4GBXSZ2O5PKSO3V4YKALANCNFSM57PFLEQA> . You are receiving this because you commented.Message ID: ***@***.***>

-- DINH Viêt Hoà