Skip to content

Bound fixed-partner two-ADD retiming across three public MOVEs - #7383

Open
ToikY wants to merge 3 commits into
CrocSwap:mainfrom
ToikY:codex/p104-three-move-two-add-retime-bound
Open

ToikY wants to merge 3 commits into
CrocSwap:mainfrom
ToikY:codex/p104-three-move-two-add-retime-bound

Conversation

@ToikY

@ToikY ToikY commented Nov 12, 2026

Copy link
Copy Markdown
Contributor

Change and mathematical scope

NO VERIFIED GAIN. The comparator is independently reproduced #7334 at aaeec8c0acf01e99a5f331ef363a698d9c084ace with signed #7113 at 07a76c5e2738882c0cb322278e56f6cae3c9885d.

This exhausts the five three-MOVE public windows in the frozen #7365 changed-role census with exactly two incident ADDs. The proposed contraction retains both original operands, their signed coefficients and order, and every other paid path. The lead tests 47,490 typed cuts over 15 direct-MOVE placements and finds zero legal ADD slots. The independent barrier walk grants any direct-MOVE clock in the original window. None of the ten partners reaches either required chart in its permitted interval. The owned donor 1555 stays in its original lifetime. All five windows are excluded within this scope.

One direct replacement would remove 34 paid positions across the signed 17 copies and leave 22,461 productive positions. At most three replacements fit after overlapping paid clocks are counted once. The optimistic compatible credit is bounded above by 1.548051774782075 raw, granting ancillary costs free. Verified saving is zero.

Task and overlap

This changes the #7365 contact-census scope to a three-MOVE contraction with two original ADDs. #7368 checks a related sole-ADD case. #7369 covers the overlapping chart-746 roles with six paid legs and fixed ADD clocks. #7371 and #7372 cover distinct two-MOVE scopes. Canonical main was e54523517d0f03e58352e647c403a5af3e525a57 at the pre-submit refresh.

Validation

Use CPython 3.12 with assertions enabled from a clean checkout.

python3.12 -I -B research/p104-three-move-two-add-retime-bound/replay.py

The Git-free replay authenticates the predecessor census and price, regenerates four matching receipts, and checks both exact audits. Synthetic ADD commutation and COPY/ownership barriers passed. In a separate relocated control, changing the census caused rejection before the audit. The research commit adds one package and uses the two existing #7257 CI fixes for the historical selected replay gap documented in #7371. The selected record is unchanged.

Credit

#7365 supplies the contact census. #7362 supplies the paid and productive rank stock. #7338 supplies overlapping two-MOVE keys and the signed admission, #7298 the directed arithmetic, and #7003 the earlier commuting slide. OpenAI Codex and subagents materially assisted.

Open next questions

Test whether different endpoint donors can supply the two incoming ADDs as one exact aggregate on the first role-532 window. The fixed-partner exclusion leaves that distinct changed-donor route open. Review and merge are left to the maintainers.

Clear1nsight and others added 3 commits November 11, 2026 15:39
This is the exact documentation prerequisite retained by CrocSwap#7156. It points the onboarding check to the selected CrocSwap#7000/CrocSwap#7001 review. Credit to Rohan Arun and the source authors.
The PR CrocSwap#6292 rejection census and package plan are statements about 2485805. Load its exact sources and planner in a private sparse checkout, keeping all source hashes, 19 expected rejections, nine package selections and the empty-plan control. Subsequent edits to the live Makefile and CI guide no longer invalidate this historical check.

The original audit is by unpaidmoon, building on Dennis’s historical replay gate. The pinned repository view follows the precedent of Matarener’s PR CrocSwap#5434. Repair prepared with substantial OpenAI Codex assistance.
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.

2 participants