fix(acir-gen): Use the side effect variable in slice_pop_back

[aakoshh]

Description

Problem

Resolves #10446

Summary

Fixes the ACIR generation of the slice_pop_back operation to use the current_side_effect_var if the length of the slice is not a constant, to calculate new_length as (slice_length - 1) * side_effects_enabled instead of just slice_length - 1.

Additional Context

The flattening pass makes the slice length a variable which is calculated based on which branch of the program was taken. In the example in the ticket, we had code like this:

if !a {
  let mut s = [10];
  if a {
    s = [20, 30];
    let (_, _) = s.pop();
  } else {
    s = [40, 50, 60, 70];
  }
} else {
  ...
}

When we flatten this for ACIR, we end up with a slice backed by an array of length 4, to accommodate the longest alternative, and a length that is a multiplication of the different length options:

• length = (1-a)*a*2 + (1-a)*(1-a)*4 + a*0 = (1-a)*4
• array = [(1-a)*a*20 + (1-a)*(1-a)*40, (1-a)*a*30 + (1-a)*(1-a)*50, 60, 70] = [(1-a)*40, (1-a)*50, 60, 70]

When a is 1, the length becomes 0, and the array becomes [0, 0, 60, 70], which is what we see in #10446 (comment)

The problem is that slice_pop_back wants to read the item at length-1, ie. the last item, which for a slice of length 0 is item -1.

With the current change, it should read item 0.

I think reading item 0 should not fail: if a slice is known to be empty, the instruction would have already been eliminated as unreachable.

User Documentation

Check one:

• No user documentation needed.
• Changes in docs/ included in this PR.
• [For Experimental Features] Changes in docs/ 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]

Changes to circuit sizes

Generated at commit: d2f70d2f63b800f396f7cc6ed49ccd964d07c255, compared to commit: e4228986ee37f02c0f55df04ec0e95d161955feb

🧾 Summary (10% most significant diffs)

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
slice_dynamic_index	+18 ❌	+2.24%	+18 ❌	+0.35%

---

Full diff report 👇

Program	ACIR opcodes (+/-)	%	Circuit size (+/-)	%
slice_dynamic_index	822 (+18)	+2.24%	5,177 (+18)	+0.35%

[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: 5bb3ce6	Previous: e422898	Ratio
test_report_zkpassport_noir-ecdsa_	2 s	1 s	2

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

CC: @TomAFrench

[jfecher]

👍