Repository navigation
Conversation
Contributor
Author
|
Hosted status for exact publication head b895ae7: planning failed, with 8 failures and2 errors only in unchanged CurrentResearchAssignmentTests (511 tests,22 platform skips). Scientific replay was SKIPPED. The existing repair is PR #5761; it is outside this focused package and remains unmerged. Registered normal/optimized and separate read-only Git-free empty-PATH cold checks pass locally. No hosted scientific success is claimed. Maintainer edits remain disabled. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Change and mathematical scope
This binds every actual G20 internal route marker to a frozen concrete normalizer word under an explicit single-ROW-cover model. The exact PR #5798 stream has 7,548 markers, with 3,774 ENTRY and RETURN records each. Every retained event index is joined to its source call, thick-stage ordinal, owner bank, port, phase and word hash.
PR #5798 already checks the complete global endpoint, marker placements and original fee; PR #5801 already constructs the 111 RIGHT normalizer/inverse words. The new result closes their stated incidence/direction gap without claiming either predecessor construction. Markers themselves do not encode a physical lowering. Under the retained convention h=gS and local P=WQW^-1, ENTRY uses inverse_word and RETURN forward_word, returning borrowed g exactly.
The source native pairing is x_i with x_(15-i). Applying W on both paired halves means physical diag(W,JWJ), not contiguous diag(W,W). An independent oracle checks 222 word products, 111 cover conjugations, 1,776 literal paired basis columns and all18 chosen layouts. Actual source ports discriminate wrong direction, missing second half and wrong reflection. Two more controls preserve aggregate counts while breaking actual bank restoration.
The selected words contain 202,640 ADDs and 73,916 SCALEs. Conditional lifting to every N=127f cover row/lane gives 25,735,280 ordered_shift and 9,387,332 unit_scale calls times f^2. One cover word per owner-marker already affects both native halves through the encoding; there is no second full ROW-word factor for those halves. Each coefficient keeps its own positive symbolic tariff. These are counts for this declared plan, not a canonical full invoice or unit prices.
Literal physical head-map correspondence, cache COPY/CLEAR, carries, low-type commit, padding, movement, restoration schedule, coefficient preparation/tariffs, compiler and supplier admission remain open. Native/global theorems remain dependencies; the five inherited framework assumptions remain. No new kappa or frontier change.
Task selection and overlap
Canonical main refreshed: de5ec2e, priority8. Source #5798: eaa9cc8. Word source #5801: eb92cbd.
Canonical claim and TASK.json explicitly narrow the formal-interface-lemma card to this particular binding audit, with no Lean build or physical implementation expansion. Live intake covered356 open PR records, actual #5777/#1166 handoffs and a focused G20 Discussion search. Sparse local inventory, Discussion search and unpublished work remain coverage limits; no-match is not novelty proof.
Exactly15 files are added in research/g20-route-word-binding. Frozen source data are authenticated by raw SHA256/Git blob pins, without predecessor implementation or cross-checker imports. Main-context ancestry and input-origin byte bindings are separate.
Validation
Python3.12.14, standard library:
Registered normal/optimized primary and separate optimized oracle pass. Their complete event-join hashes and every coefficient count agree. Seven primary controls rebuild the altered bill before semantic rejection where applicable; the oracle adds five actual algebra/owner/restoration controls. Complete canonical receipts preserve numeric-type distinctions.
Committed bytes/modes/declared main history and official export checks pass. A read-only Git-free isolated optimized replay with empty PATH passes both checkers in about3.53 seconds, preserving all16 exported files including the integrity manifest. Export SHA256: e7ed8b63dd3f890ff0216d66e5fcbb403f0d73093e5203625ba6553c8bed5a6e. Current selected baseline passes bindings/metadata only.
Publication head b895ae7; tree84f7450f477c985e5bd2571bf74181d1b254b682. Hosted checks not yet observed; no hosted scientific pass is claimed. The shared assignment-suite repair remains #5761.
Credit and next obligation
Ali Romman with Codex assistance supplied the G20 construction/audits and normalizer witnesses. Retain Dugongue, ai-nanachi, #5679/#5708/#5782 and all source lineage. JoyboyBrian supplied the earlier ROW/P2 schedule and its retained predecessors. ToikY/Codex #5801 words are frozen predecessor work. New marker-to-word binding, conditional cover convention and reflected-address audit: ToikY with substantial OpenAI Codex assistance, separate primary/oracle/proof-review contexts. Classical conjugation is not a new general theorem. Apache-2.0; NOTICE retains attribution.
Next: bind one actual marker to the literal physical ROW head/cache/carry/commit/padding/return implementation and prove coefficient-specific tariffs before importing this plan into a full caller invoice. Independent acceptance remains. Maintainer edits disabled; no merge or closure requested.