Add recursive-tree e2e goldens over the leaf bootloader

[YairVaknin-starkware]

Add goldens at test_data/goldens/four_leaves/: a real leaf (the leaf simple bootloader running a simple-output task, proven end-to-end by leaf_prover, wrapped into a LeafInput with the hashed-output preimage injected the way the backend does) duplicated 4x and folded. supported_preprocessed_roots.json is the unpacker's trust-anchor list (the circuit-world analogue of the supported program hash lists). The committed leaf_simple_bootloader_compiled.json (landed in #391) is compiled from the main starkware repo revision that moves run_simple_bootloader_single_task into run_simple_bootloader.cairo (the goldens are proven against it).

The new test_golden_four_leaves_e2e test is the true end-to-end AND the goldens generator: it runs leaf_prover itself (via prove_leaf_from_files), folds 4 copies, and asserts both the independently recomputed hash tree (every hash layer, from the raw preimage up) and equality with the committed goldens; run with FIX=1 to regenerate them (see the e2e module docs). The fold e2e tests switch from the interim leaf_fixture.json (same content, removed here) to the golden leaf.

Stack (4/4): prove_leaf_from_files (#390) ← retarget canonical (#391) ← PackedNode roots (#392) ← this.

🤖 Generated with Claude Code

---

This change is 

[cursor]

PR Summary

Low Risk
Changes are confined to test fixtures, dev-dependencies, and slow-test documentation; production fold/prove binaries are unchanged.

Overview
Adds committed regression goldens under test_data/goldens/four_leaves/ (leaf proof, preimage dump, 4-leaf fold outputs, manifest, and supported_preprocessed_roots.json) and wires slow e2e tests to that tree instead of the removed interim leaf_fixture.json.

test_golden_four_leaves_e2e runs leaf_prover on the leaf simple bootloader + simple-output task, injects the hashed-output preimage like the backend (hex dump → decimal felts via num-bigint), folds four copies, and asserts byte-for-byte JSON equality with the goldens; FIX=1 on that test refreshes all golden files including the trust-anchor roots list. The lighter test_fold_two_leaves / test_fold_three_leaves_with_carry cases now duplicate goldens/four_leaves/leaf.json while still independently recomputing the packed hash tree.

dupe_and_fold returns the multiverifier preprocessed root so the generator can write supported_preprocessed_roots.json.

^{Reviewed by Cursor Bugbot for commit f2839a4. Bugbot is set up for automated code reviews on this repo. Configure here.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 65.88%. Comparing base (ae52f97) to head (f2839a4).

Additional details and impacted files

@@                           Coverage Diff                           @@
##           yairv/packed_node_preprocessed_root     #388      +/-   ##
=======================================================================
+ Coverage                                65.59%   65.88%   +0.28%     
=======================================================================
  Files                                       45       45              
  Lines                                     5906     5906              
=======================================================================
+ Hits                                      3874     3891      +17     
+ Misses                                    2032     2015      -17     

see 2 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[OmriEshhar1]

@OmriEshhar1 made 1 comment.
Reviewable status: 0 of 21 files reviewed, 1 unresolved discussion (waiting on Gali-StarkWare and YairVaknin-starkware).

---

a discussion (no related file):
it looks like there are at least 3 separate changes here:

1. refactor in leaf_prover (prove_leaf_from_files)
2. changes to canonical
3. 'golden' test
   Consider separating to 3 (or more) more manageable PRs.

[YairVaknin-starkware]

@YairVaknin-starkware made 1 comment.
Reviewable status: 0 of 22 files reviewed, 1 unresolved discussion (waiting on Gali-StarkWare and OmriEshhar1).

---

a discussion (no related file):

Previously, OmriEshhar1 wrote…

it looks like there are at least 3 separate changes here:

1. refactor in leaf_prover (prove_leaf_from_files)
2. changes to canonical
3. 'golden' test
   Consider separating to 3 (or more) more manageable PRs.

Done.

[OmriEshhar1]

@OmriEshhar1 reviewed 22 files and all commit messages, and resolved 1 discussion.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on Gali-StarkWare).

[OmriEshhar1]

@OmriEshhar1 reviewed 12 files and all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on Gali-StarkWare).

[Gali-StarkWare]

@Gali-StarkWare reviewed 22 files and all commit messages, and made 1 comment.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on YairVaknin-starkware).

---

crates/stwo_run_and_prove_recursive_tree/src/tests.rs line 211 at r5 (raw file):

    }

    /// The golden `leaf_prover` output the cheaper e2e folds duplicate their leaves from: the leaf

I didn't know we're in ancient greece :)

Code quote:

The golden `leaf_prover`