Tracking issue: Fix remaining failing tests in the SSA Refactor

[jfecher]

Aim

All tests in the old test_data folder should also compile with the SSA refactor.

Expected Behavior

When compiling each test in test_data with the --experimental-ssa flag, they should all prove and verify.

Bug

Instead, the following tests are still failing with the following reasons:

• 9_conditional: divide by zero while proving (Divide by zero in 9_conditional #1802)
• array_dynamic: expected array index to be a known constant, requires dynamic array support in ACIR
• bit_shifts_runtime: (non-blocking) not implemented: rhs must be a constant when doing a right shift, requires dynamic shift support in ACIR
• eddsa: hangs - EDIT: does not hang, just takes 16 minutes to complete...
• higher_order_functions: lookup: variable not defined, requires closure conversion, see Implement closure conversion #1254, and Closures are unsupported in the ssa refactor #1662
• regression: Error: could not satisfy all constraints (Could not satisfy all constraints for regression test #1800)

Please keep this list updated by checking each test once each bug is fixed and the test starts passing again. When all tests are passing we can close this issue.

[Savio-Sou]

Thanks Jake for creating the issue! Is the list of bugs up-to-date? cc @kevaundray

P.S. feel free to make use of GitHub Tasklist if preferred with the syntax:

> ```[tasklist]
> ### Tasks
> - [ ] Add a draft title or issue reference here
> ```

(Omit the ">", they're just there for formatting.)

[dbanks12]

Watching this issue as it is blocking the "spending pending commitments" feature in Aztec. Wondering if this is quite a bit of work to do and what the timeline is for it?

[jfecher]

@dbanks12 this is a tracking issue for several smaller issues. Do you know which specifically is blocking?

[TomAFrench]

@jfecher All of them are 😂. David is wanting us to switch over to the new SSA so we can start using the stepwise ACVM.

[kevaundray]

@jfecher All of them are 😂. David is wanting us to switch over to the new SSA so we can start using the stepwise ACVM.

It would be:

• 9_conditional (Kev)
• array_dynamic (Guillaume)
• bit_shifts_runtime (Guillaume)
• regression (Kev)
• eddsa (Maxim/Kev)

With the ones in bold being the ones that are needed

[TomAFrench]

bit_shifts_runtime

This shouldn't be a blocker for switching over as this is unimplemented in the old SSA as well afaik.

[jfecher]

I don't think eddsa should be a blocker either. It looks like it doesn't actually hang, but just takes 16 minutes to prove.

[kevaundray]

Closing this issue, the two remaining opened issues are not affected by this tracking issue. To track the status of bit shift runtime, see here