chore(ssa refactor): Simplify inlining pass and fix inlining failure

[jfecher]

Related issue(s)

Resolves #

Description

Summary of changes

This simplifies the inlining pass by removing the extra block that is inserted for return instructions to jump to. Instead, when a function call is inlined we just switch back to the same block as the return instruction. The result is fewer blocks created which allows the code in issue #1327 to compile properly in the ssa refactoring without needing a block simplification pass afterward and subsequent second inlining.

Dependency additions / changes

Test additions / changes

I've added a unit test for the issue in #1327, although the issue is still open since the experimental ssa refactor is not ready to be the default IR.

Checklist

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt with default settings.
• I have linked this PR to the issue(s) that it resolves.
• I have reviewed the changes on GitHub, line by line.
• I have ensured all changes are covered in the description.

Documentation needs

• This PR requires documentation updates when merged.

Additional context

[jfecher]

To elaborate, previously the code in that issue would compile to (after inlining):

fn d_method f1 {
  b0(v0: Field):
    v2 = mul v0, v0
    return v2
}
fn main f4 {
  b0(v0: Field):
    jmp b2(f1)
  b2(v3: function):
    jmp b3(v3)
  b3(v5: function):
    v6 = call v5(v0)
    return v6
}

Note the extra blocks in main. The function being passed through block parameters made it so inlining couldn't see what function was referred to by v5 and failed to inline the final call. With this PR the extra blocks created from function calls are eliminated and the new IR is:

fn main f4 {
  b0(v0: Field):
    v4 = mul v0, v0
    return v4
}