fix: Improve floating point lint to handle ambiguous type

[odysa]

changelog: [suboptimal_flops]: improve handling of ambiguous float types in mul_add suggestions

Fixes: #14897, #10015

This is not a perfect solution, but it fixes the issue. Happy to discuss further with the clippy team.

This PR refines check_mul_add to avoid false positives on ambiguous float types (e.g., f32 vs f64). It improves type detection and ensures mul_add is only suggested when the type is clear.

Changes:

• Improved detection of ambiguous float types
• Refined mul_add suggestion logic
• Added tests for ambiguous cases

Example

Before this change, expressions like:

let x = 1.0; // ambiguous type
let _ = x * 2.0 + 0.5; // could incorrectly suggest mul_add

[rustbot]

r? @llogiq

rustbot has assigned @llogiq.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer