Fix KeyMaterial::concatenate logic and improve Condition<T> correctness for i64/u64

[Quant-TheodoreFelix]

Title

Fix KeyMaterial::concatenate logic and improve Condition<T> correctness for i64/u64

Description

This PR addresses three main issues: a logical bug in KeyMaterial::concatenate regarding security strength, the missing implementation of Condition<u64>, and a correctness fix for the Condition<i64>::TRUE constant.

1. Fix func concatenate in KeyMaterial Security Logic

• Issue: The documentation for KeyMaterial::concatenate states that combining a full entropy key with a low entropy key should result in a low entropy key (conserving the weaker security parameters). However, the implementation was using max(), which effectively "upgraded" the security strength of the resulting key to the higher of the two inputs.
• Fix: Changed the logic to use min() for both key_type and security_strength to strictly adhere to the documentation and safe security modeling principles.

2. Implement u64 Condition

• Issue: Condition<u64> was defined but lacked a concrete implementation, preventing constant-time operations on u64 primitives.
• Fix: Added impl Condition<u64> including:
  • Correct mask generation in from_bool using 0u64.wrapping_sub(value).
  • Correct TRUE constant definition as u64::MAX (all-ones) to ensure bitwise operations like select work correctly across all bits.
  • Implementation of select, is_true, and other helper methods.

3. Fix TRUE Constant in i64 Condition

• Issue: Condition<i64>::TRUE was previously defined as 1. This caused select and negate functions (which rely on bitwise masking val & mask) to fail for all bits except the LSB.
• Fix: Updated Condition<i64>::TRUE to -1 (all ones in two's complement) to ensure the mask covers the entire integer width.

4. Tests

• Added a new test module u64_tests in ct_tests.rs to verify Condition<u64> behavior.
• Added a specific test case for or_halves to verify bit-folding logic.
• Verified that select and negate logic behaves correctly with the updated TRUE constants.

Motivation

To ensure the library adheres to its stated security contract regarding key material degradation and to provide correct, functional constant-time primitives for 64-bit integers.

Checklist

• Code compiles correctly.
• Added tests for Condition<u64> and or_halves.
• cargo test passes.
• Documentation comments updated/verified.