[WIP] Chore: leaky decimal fixes not implemented

[connortsui20]

DO NOT MERGE YET

I just had these changes staged on my laptop so I don't want to lose them.

At some point we need to create an issue for the decimal logical and physical type separation issue (maybe I will get around to that soon).

Part of these changes should also fix all of the fuzzer issues being titled incorrectly (a bunch of null fuzzer issues that are really cast errors).

[codspeed-hq]

CodSpeed Performance Report

Merging #5817 will improve performance by 39.38%

_{Comparing ct/leaky-decimals (8a8b7c1) with develop (c7b0ef9)}

Summary

⚡ 21 improvements
✅ 1233 untouched
⏩ 623 skipped¹

Benchmarks breakdown

	Benchmark	BASE	HEAD	Efficiency
⚡	new_bp_prim_test_between[i16, 32768]	141.9 µs	112.1 µs	+26.57%
⚡	new_bp_prim_test_between[i16, 16384]	86.7 µs	71.9 µs	+20.57%
⚡	new_bp_prim_test_between[i64, 16384]	137.3 µs	121.2 µs	+13.28%
⚡	new_bp_prim_test_between[i32, 32768]	175.4 µs	145.5 µs	+20.54%
⚡	new_bp_prim_test_between[i64, 32768]	241 µs	208.5 µs	+15.6%
⚡	new_bp_prim_test_between[i32, 16384]	104.3 µs	89.4 µs	+16.65%
⚡	new_raw_prim_test_between[f32, 32768]	163.5 µs	117.3 µs	+39.38%
⚡	new_raw_prim_test_between[i32, 16384]	80.9 µs	66 µs	+22.45%
⚡	new_raw_prim_test_between[f64, 16384]	121.5 µs	97.4 µs	+24.68%
⚡	new_raw_prim_test_between[i64, 32768]	197.9 µs	165.4 µs	+19.67%
⚡	new_raw_prim_test_between[f64, 32768]	221.1 µs	172.7 µs	+28.01%
⚡	new_raw_prim_test_between[f32, 16384]	92.9 µs	69.9 µs	+32.89%
⚡	new_raw_prim_test_between[u32, 32768]	142.2 µs	112.8 µs	+26.11%
⚡	new_raw_prim_test_between[u32, 16384]	81.7 µs	67.2 µs	+21.62%
⚡	new_raw_prim_test_between[i64, 16384]	109.8 µs	93.7 µs	+17.18%
⚡	new_raw_prim_test_between[i32, 32768]	140.1 µs	110.3 µs	+27.05%
⚡	new_raw_prim_test_between[u64, 16384]	110.2 µs	93.8 µs	+17.38%
⚡	new_raw_prim_test_between[u64, 32768]	198.4 µs	165.6 µs	+19.81%
⚡	new_alp_prim_test_between[f64, 16384]	155.6 µs	139.5 µs	+11.49%
⚡	new_alp_prim_test_between[f64, 32768]	210.2 µs	177.8 µs	+18.25%
...	...	...	...	...

ℹ️ Only the first 20 benchmarks are displayed. Go to the app to view all benchmarks.

Footnotes

1. 623 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[connortsui20]

related: #5820