[1337isnot1337]

Mixing two original P112 coordinate terms with one nonzero free odd rectangle through GL(3,C) cannot improve the COPY bill when all three replacement terms retain original-coordinate components on both sides. For the fixed pair (4,6), each term costs at least six; the whole decoder costs at least 66 versus 64. If both projected coordinates are present in every factor, the whole floor is 72. At unchanged width/normalization and 0<q=1-a<=1, the full COPY power bill is at least 33/32 of the baseline, even granting all added preparation, source access, frame, mixing, restoration, hosting and child-management expenses zero.

Scoped negative result; no new kappa or physical supplier. The result leaves the exactly-two projected-product branch open. The identity sum q_j p_j^T=I_2 proves that at least two products must be active, so it identifies the remaining algebraic boundary for this mechanism. Zero rectangles, other coordinate pairs and changed normalization are outside this theorem.

The proof reduces arbitrary complex original-coordinate normals to five exhaustive support classes. Two independently implemented exact engines cover all 25 class pairs and all 5,824 ordered odd port pairs: 145,600 records, with refined survivor minima 6/7/8. Their cost bytes are compared directly, alongside every histogram, witness and slice dimension. A support-containment argument then handles arbitrary nonzero free rectangles and complex coefficients. Four Q(i) fixtures check all 12,544 entries each. Two concrete (8,8,0) boundary identities demonstrate why the hypotheses cannot be removed.

OS assignment seed 46, structural decoder track. Exact base: bce54fb3f895cbe5cd4c0eb477c3afeed100190f; selected conditional kappa there is 0.001672660197076918, PR1094. Actual model input is retained PR845 geometry, derived from PR717 at 4150829a2cb66b823f4282da98985959ee448a90. Exact SHA256/Git blob pins, pending-source heads and concurrent-claim distinctions are included.

Reproduce:

python3 -I -B research/p112-gl3-survivors-seed46/verify.py

Expected: PASS_P112_GL3_S46 records=145600 floor=66 power_ratio_floor=9/8 controls=10 no_new_kappa.

Passed direct certificate replay, registered metadata/replay, all ten rejection controls, an optimized CLI receipt-corruption test, selected-record integrity and whitespace checks. The complete overlap refresh covered 838 open PRs and 88 discussion/comment/reply records, followed by a final recent-activity check. No full supplier, native or hosted-CI success is claimed. Separate-total COPY-tree and full compiler/frame/hosting/precision/analytic obligations remain inherited.

Credit Arham Tahir's PR1331, head f08b698cc2bd74e2425486728a008e531fc35f0e, for the one-coordinate GL2 predecessor and survivor-containment method; PR845/Mihir Srivastava, PR717/446/DreamingOfClouds, PR620/GamingPuzzled and the retained pairing-cost lineage.

Concurrent GL3 singleton work by commdata2338 aims at all branches and all 28 coordinate pairs. Arham Tahir's rank-nine singleton claim allows all nine terms to change. This package adds arbitrary-rectangle coverage for one pair's three-product branch and a reproducible 6/7/8 survivor table. Those active claims are credited, not assumed proved or used as theorem dependencies.

Prepared by Cinna Davis (@1337isnot1337), with substantial OpenAI Codex assistance in research, proof, implementation, verification and documentation; Apache-2.0; no independent human review or formal multiplication theorem claimed. Only one new scoped research package is added, with concrete open next questions.

[Mister-Meeseeks]

100-PR batch review at 3eaef3e4427068b33847dc1147c8a2e5709a970e.

Merged to main, preserving the original contributor commits and package attribution.

Reviewed scope: Scoped GL(3,C) mixing obstruction for two P112 coordinate terms plus one nonzero free rectangle. Complete support-stratum pairing census and survivor argument give COPY at least 66 in the stated nonzero-projection class; zero-factor and larger-factorization cases are distinguished.

The pinned scoped command passed; the batch ledger contains its command, input hashes, environment and log. This incorporates the research/tooling contribution and does not change the selected κ or discharge the inherited multiplication-theorem assumptions.

Batch review, attribution and evidence. Thank you for the contribution.