Replace once_cell with std Implementation

[uulm-janbaudisch]

The dependency on once_cell is not needed any more since the standard library implements its functionality.

Additionally replace a Mutex with RwLock since read / write usage is clearly separated.

Depends on #104

[github-actions]

Bencher Report

Branch	rm-once-cell
Testbed	aarch64-linux

🚨 1 Alert

Benchmark	Measure Units	View	Benchmark Result (Result Δ%)	Upper Boundary (Limit %)
load auto1 (d4)	Latency milliseconds (ms)	📈 plot 🚷 threshold 🚨 alert (🔔)	41.10 ms (+4.09%) Baseline: 39.49 ms	40.54 ms (101.40%)

Click to view all benchmark results

Benchmark	Latency	Benchmark Result microseconds (µs) (Result Δ%)	Upper Boundary microseconds (µs) (Limit %)
load auto1 (d4)	📈 view plot 🚷 view threshold 🚨 view alert (🔔)	41,105.00 µs (+4.09%) Baseline: 39,491.67 µs	40,538.21 µs (101.40%)
load auto2 (c2d)	📈 view plot 🚷 view threshold	565,160.00 µs (+5.30%) Baseline: 536,705.00 µs	567,584.60 µs (99.57%)
load busybox (c2d)	📈 view plot 🚷 view threshold	1,426.20 µs (-0.74%) Baseline: 1,436.89 µs	1,460.56 µs (97.65%)
load small (c2d)	📈 view plot 🚷 view threshold	5.76 µs (-0.65%) Baseline: 5.80 µs	5.92 µs (97.23%)
t-wise busybox (c2d) t=2	📈 view plot 🚷 view threshold	975,100.00 µs (+14.78%) Baseline: 849,501.67 µs	1,030,002.75 µs (94.67%)
t-wise small (c2d) t=2	📈 view plot 🚷 view threshold	11.18 µs (-5.35%) Baseline: 11.81 µs	12.60 µs (88.72%)
t-wise small (c2d) t=3	📈 view plot 🚷 view threshold	12.51 µs (-3.78%) Baseline: 13.00 µs	13.70 µs (91.32%)

🐰 View full continuous benchmarking report in Bencher