Skip to content

Certify the A1858 contraction's cleanup edit. - #4152

Merged
Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
unpaidmoon:cursor/a1858-cleanup-contraction-3e41
Oct 20, 2026
Merged

Mister-Meeseeks merged 1 commit into
CrocSwap:mainfrom
unpaidmoon:cursor/a1858-cleanup-contraction-3e41

Conversation

@unpaidmoon

@unpaidmoon unpaidmoon commented Oct 20, 2026 •

Copy link
Copy Markdown
Contributor

Change and mathematical scope

Partial audit, no new κ. On the four frozen PR #3998 words, the consecutive A1858 triple and its two-ADD image have the same exact rational matrix. An independent cleanup derivation reproduces each stored cleanup section. The contraction replaces one three-gate inverse block by the two-gate inverse of the image: one ADD is deleted and the two survivors exchange order. Forward and cleanup ADD counts each fall by one. The selected frontier is unchanged.

Task and overlap

Base bb4bccbf42844d2801487450727c2f63e10d1d0f. Predecessor head 389268a6b5e0373a3125d5c6fcd6995f4ea324f5 (PR #3998). Card independent-moment, scope refined away from the #2979 rank-7 reconstruction.

This is item 7 of the chronological open #1166 next-task list after dropping work delivered or refined by open PRs through #4149. The question is discussion comment 19035159: challenge the source-bound A1858 contraction and its cleanup. It is not PR #4019's separated seven-block contraction, which #4112 already accepts.

Validation

python3 scripts/research_package.py --package research/staged544-a1858-contraction-audit prints PASS_A1858_CLEANUP_CONTRACTION and PASS_REGISTERED_RESEARCH_PACKAGE ... outcome=partial. CPython 3.12, standard library only. Controls reject a wrong central sign, a reversed cleanup order, a helper ADD omitted from cleanup, and a target ADD inserted into cleanup.

Not run: the frame-bill replay, native/owner/precision obligations, and any supplier moment.

Credit

Dataisfire (PR #3734), Dugongue (PR #3945), and Chris R / Clear1nsight (PR #3998 and comment 19035159). unpaidmoon, with substantial Cursor Grok assistance. Apache-2.0.

Open next questions

Recompute the successor climb of registers 456, 457 and 475 before treating the published histogram as unchanged. The second half of comment 19035159, carrying the scalar saving into the PR #3945 producer, remains open.

Independently derive the one-ADD cleanup substitution and the surviving
inverse order on the four frozen PR3998 words. No new kappa.

Co-authored-by: unpaidmoon <unpaidmoon@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants