fix: incorrect max bit size in remove_bit_shifts

[asterite]

Description

Problem

Resolves #9541

Summary

When decomposing an exponent into bits we always need to take into account the maximum number of bits the exponent type can hold. Previously if the exponent was casted from a smaller type, for example from u8 to u32, we'd use the maximum number of bits of values in u8, which is 8, then take ilog2 of that, which is 3, then add 1 to get 4. But values that fit in u8 that are less than 32, like 26, need more than 4 bits to be represented. Given that the maximum value to shift an u32 is 32, it makes sense that the maximum bits is the bits to represent 32.

Additional Context

Documentation

Check one:

• No documentation needed.
• Documentation included in this PR.
• [For Experimental Features] Documentation to be submitted in a separate PR.

PR Checklist

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[github-actions]

Changes to circuit sizes

Generated at commit: 87e0b6668b229f9ccc76d5bccbb9641541c7e651, compared to commit: 52a8579901ba03c53cbb022d968435bb4f66e2c8

🧾 Summary (10% most significant diffs)

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
u16_support	-4 ✅	-10.53%	-5 ✅	-0.18%
bit_shifts_runtime	-24 ✅	-4.41%	-35 ✅	-0.72%

---

Full diff report 👇

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
u128_type	134 (-4)	-2.90%	4,617 (-5)	-0.11%
binary_operator_overloading	88 (-4)	-4.35%	3,300 (-5)	-0.15%
inactive_signed_bitshift	76 (-4)	-5.00%	2,844 (-5)	-0.18%
u16_support	34 (-4)	-10.53%	2,792 (-5)	-0.18%
bit_shifts_runtime	520 (-24)	-4.41%	4,860 (-35)	-0.72%

[github-actions]

⚠️ Performance Alert ⚠️

Possible performance regression was detected for benchmark 'Test Suite Duration'.
Benchmark result of this commit is worse than the previous benchmark result exceeding threshold 1.20.

Benchmark suite	Current: 3ea2e92	Previous: 910ed69	Ratio
test_report_zkpassport_noir-ecdsa_	2 s	1 s	2

This comment was automatically generated by workflow using github-action-benchmark.

CC: @TomAFrench