fix: bind self generic type in trait calls via a concrete type

[asterite]

Description

Problem

Resolves #8794

Summary

The main issue here was that when you had a call like this:

SomeConcreteType::trait_method(...)

that would resolve to a trait method, but we'd throw away the information that this was called on SomeConcreteType.

Additional Context

So in the example of #8794:

• when we elaborate MyU128::from we end up with a type fn('4) -> '5
• then when we elaborate the call, MyU128::from(x), given x : u64 we end up with fn(u64) -> '5

The problem is that the return type is still unknown at this point.

Checking with Rust, it could figure out the return type even if we didn't call the function, for example like this:

// Rust doesn't know the argument types, but it knows it returns MyU128
let f = MyU128::from;

We already have specific code to solve SomeConcreteType::trait_method as a reference to a trait method, but the information about the specific type, SomeConcreteType, was thrown away. Now it's captured, then when elaborating it we'd bind the "Self" generic type of a function (if there's any) to that type.

Or at least this is what I felt was missing. Let me know if this should be done some other way.

Now it behaves like Rust, so we can actually do this:

// Noir now knows f returns MyU128
let f = MyU128::from;
let _ = f(x);

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]

⚠️ 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: a996045	Previous: 8598eab	Ratio
test_report_noir-lang_noir_bigcurve_	287 s	234 s	1.23

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

CC: @TomAFrench

[TomAFrench]

LGTM