test(flatten_cfg): unit test and end to end test for #1792

[ludamad]

Description

Problem*

Resolves

Summary*

Documentation

• This PR requires documentation updates when merged.

  • I will submit a noir-lang/docs PR.
  • I will request for and support Dev Rel's help in documenting this PR.

Additional Context

PR Checklist*

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

[ludamad]

GPT4 did the conversion after given one example, plus a bit of hand-massaging

[ludamad]

Simplified (by GPT4 of course I'm not a sucker)

[ludamad]

@kevaundray @jfecher figured out that the simplify code thinks

        // fn main f1 {
        //     ... simplified ...
        //     v_false = u1 0
        //     jmpif v_false then: b_useless_branch, else: b_fallthrough
        //   b_useless_branch():
        //     jmp b_fallthrough()
        //   b_fallthrough():
        //     constrain v_false // was incorrectly removed
        //     return
        // }

the constrain here can be removed on flatten_cfg, but I didn't debug further.

[jfecher]

Can we open this as an issue in the future? Unless the PR already has the fix for the bug included I think it is better to file the bug first.

[ludamad]

Sure, although surely this needs the fix any way and then it has an issue? This can't be merged without it. I thought of this more as a hand off PR

[jfecher]

Sure, although surely this needs the fix any way and then it has an issue? This can't be merged without it. I thought of this more as a hand off PR

Sorry, I misinterpreted this PR at first. I didn't see the existing linked issue #1792 somehow. I think for cases like this we can include any testing code in the issue and just make sure to add it in the same PR the issue is fixed in. That way we have assurance the PR fixes the issue in question compared to adding the test afterward.

[ludamad]

Really I should have probably slapped DO NOT MERGE on this, it was meant to be used as you did, cherry-picked into the fix PR. And it is in the issue by virtue of mentioning the issue, but I'll put it in the body in the future