chore: do not differentiate fixed-length vs variable-length in Poseidon2

[guipublic]

Description

Problem*

We do not differentiate anymore fixed-length and variable-length poseidon2 hashes.

Summary*

Simply remove the terminator that was added at the end of the input.

Additional Context

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]

Changes to Brillig bytecode sizes

Generated at commit: dabed2c2ef62f6b9b61476a28e1699baa2e561be, compared to commit: 3a42eb5c68f9616f0ebe367c894f0376ba41e0ef

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
no_predicates_numeric_generic_poseidon_inliner_min	-31 ✅	-4.84%
fold_numeric_generic_poseidon_inliner_min	-31 ✅	-4.84%
fold_2_to_17_inliner_zero	-89 ✅	-22.14%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
regression_5252_inliner_min	3,572 (-16)	-0.45%
bench_2_to_17_inliner_min	355 (-15)	-4.05%
poseidon2_inliner_min	362 (-16)	-4.23%
fold_2_to_17_inliner_min	383 (-19)	-4.73%
no_predicates_numeric_generic_poseidon_inliner_min	610 (-31)	-4.84%
fold_numeric_generic_poseidon_inliner_min	610 (-31)	-4.84%
fold_2_to_17_inliner_zero	313 (-89)	-22.14%

[github-actions]

Changes to number of Brillig opcodes executed

Generated at commit: dabed2c2ef62f6b9b61476a28e1699baa2e561be, compared to commit: 3a42eb5c68f9616f0ebe367c894f0376ba41e0ef

🧾 Summary (10% most significant diffs)

Program	Brillig opcodes (+/-)	%
poseidon2_inliner_min	-4 ✅	-0.46%
fold_2_to_17_inliner_zero	-170,825 ✅	-12.96%

---

Full diff report 👇

Program	Brillig opcodes (+/-)	%
bench_2_to_17_inliner_min	712,412 (-3)	-0.00%
fold_2_to_17_inliner_min	1,355,048 (-9)	-0.00%
regression_5252_inliner_min	1,032,592 (-7)	-0.00%
fold_numeric_generic_poseidon_inliner_min	5,851 (-7)	-0.12%
no_predicates_numeric_generic_poseidon_inliner_min	5,851 (-7)	-0.12%
poseidon2_inliner_min	869 (-4)	-0.46%
fold_2_to_17_inliner_zero	1,146,797 (-170,825)	-12.96%

[TomAFrench]

I think we can hold off from making this change until we have Poseidon2 out in an external library

[guipublic]

I think we can hold off from making this change until we have Poseidon2 out in an external library

I don't see any benefit in holding this off?

[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: 1032e6e	Previous: 410c8b0	Ratio
AztecProtocol_aztec-packages_noir-projects_noir-protocol-circuits_crates_blob	62 s	50 s	1.24

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

CC: @TomAFrench

[saleel]

@TomAFrench someone in discord was complaining about this issue.
Isn't it better to have this fixed now to not have a discrepancy when its moved to a separate lib anyway?

[lucasxia01]

I thought we were keeping a variable length implementation? like with message_size being a witness value

[guipublic]

I thought we were keeping a variable length implementation? like with message_size being a witness value

Yes we still do this, it's just the 0/1 at the end that we don't do anymore.

[lucasxia01]

looks good to me