Exclude two P112 profiles by five-endpoint A4 shortening

[petalas]

Summary

9 independent pieces, code intersection dimension >= 6
  delete two unrestricted pieces
7 independent pieces, code intersection dimension >= 4
  five-endpoint A4 shortening
required endpoint support 106 or 114 > allowance 100

Prove a characteristic-zero P112 shortening contract with five support-twenty endpoints and common small support at most 26. Its general contradiction is 102>100. Apply it to the actual historical TWOSIDE (6,3) profiles 3,6^6,7^2 and 4,5,6^5,7^2.

One isolated package, research/p112-a4-shortening/, contains the proof, exact receipt, source pins, replay and controls. Commit 66ce40a4bf1c82a0e8b6e94067dc0ab2fbb4c008; canonical-main baseline 463113d675af7410ac91d53a4fed51ec6494eb9a, priority revision 6.

Credit DreamingOfClouds/PR3154 at 708f0ad8928c383de8294aea3c126e9461d2066d for TWOSIDE and its retained E8/P112 lineage, Nick Petalas/PR3164/3173 for geometry, PR3175 for the stabilizer-spectrum method, and PR3195/3197 for earlier deletion applications. PR3203/3205's six-endpoint results are distinct and credited. Development used substantial automated assistance in drafting and checks; no independent human review is claimed.

Evidence

• Before: both rows survive the original seventeen-profile (6,3) filters. Their weight-certified original covers need four/eight paid cuts, outside the earlier two-cut rule; six-endpoint results do not directly certify the four-dimensional retained slice.
  After: the new A4 contract gives 106>100 and 114>100. Full C5 and kappa are not claimed.
• Registered CPython 3.12.15 replay passes all A4 extension orbits, 4,097,782 five-/six-flats, 7,140 root planes and ten optimized behavioral groups. A separate root/cost/deletion oracle agrees.
• The exact committed, read-only, Git-free export passes both commands under optimized CPython 3.13.5. All thirteen committed package/input byte comparisons pass. Ruff, formatting, warnings-as-errors compilation and source-origin checks pass.
• Fresh overlap covered 255 open PRs, 2,630 closed/merged PRs and 2,020 Discussion records, with complete pagination and no warnings. Exact-target matches were only this local package. This is not proof of novelty or a survey of every historical implication.

python3.12 -B scripts/research_package.py --package research/p112-a4-shortening

Merge Danger

Door: two-way. Blast radius: isolated.

Only a new opt-in package is added. The standard Weyl-chamber theorem and original decoder reduction remain dependencies. This is author-checked proof and finite evidence, not a complete decoder floor, physical supplier, multiplication theorem or new kappa. The support27 relaxation reaches equality and is not a counterexample. Selected results and frozen inputs are unchanged; hosted CI is separate from local validation.