Repository navigation
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This construction adds charged bank selection and complete head return to the private reset service in PR #7663. It proves an all-size local cost and preservation theorem at one source-bound owned MOVE. No new kappa, complete recursive allocator or discharged framework assumption is claimed.
Read the proof, checked receipt, two-head controller, verifier, copied source pins and retained proof input. This answers a narrow part of the released guide/view handoff.
The exact source is
b81f40997de646de30749d367484d89e59fc3013, reconstructed wordf95eabd1734c7a74f58428cb12f261a81032bef01aed78c8147a4e4f90636939. At clock 88263, owned role 1510 moves from chart 477 to chart 1090 during lifetime ordinal 35; 34 parent lifetimes are live. A direct binary span check confirms a rank-four nested boundary. After the retained native adapters put the view in canonical order, parent shape(R,9f,W)and child shape(R*2^(5f),4f,W)have identical complete records and cell order. The proof gives the inverse address map; equal extent alone is insufficient.The finite selector reads the immutable guide and moves two distinct heads one cell at a time. It then runs the unchanged source reset graph and returns to the global origin. For bank k with L payload cells, the exact charge is
8k(L+2)+9L+14, counting guide reads, data writes, both heads' unit moves and executed control rows. Every unselected cell, guard and guide cell stays unchanged. All historical banks remain resident.For P parent slots and C distinct child slots, one INIT and CLEAR per child cost
C(16P+8C+10)L+C(32P+16C+12). The source-bound P=C=234 fixture gives1316484L+2631096. Data and guide capacity is2(P+C)(L+2)cells, plus positive control and preparation costs. The 9,398 reusable service rows have a separate constructive bound for a full static dispatcher. Guide construction, numerical child work, adapters and cross-depth allocation are not included in this local charge.Use Python 3.12 and the standard library:
Expected output includes:
Normal and optimized isolated replay passed. The checks cover all 468 source INIT/CLEAR services at L=5, three shape maps over 263,680 addresses, and a complete L=2560 fixture with 1,199,016 cells on each tape. The latter executes INIT and CLEAR at the first and last child slots and checks every unselected cell. A confined write-only borrower changes parent slot 35 while preserving the rest. Five controls detect wrong same-size format, missing spectators, ownership aliasing, short extent and omitted head return. Committed bytes, source ancestry and official export passed. Read-only Git-free optimized replay passed with source hashes and modes unchanged. Hosted CI did not certify this package. The plan job failed in the shared
test_current_ci_section_matches_selected_record: the base guide names frontier6030 while the selected record requiresmake frontier7000-verify. The log reports 760 tests, one failure and 24 skipped; package replay was skipped. This package changes no shared CI file. Local replay remains the scientific evidence.Limits: the numerical child is an explicit contract. The write-only borrower is a frame-condition witness, not a Gaussian child. The 234-slot child tuple models one raw native lifetime plan; it is not the actual inventory of a complete compiled child. Correct complete guides, distinct slots, common root width and canonical post-adapter order are premises. The actual compiled owner map, full guide production and retirement, numerical child, adapter movement, auxiliary stock and whole physical tariff remain open. The full source application and selected multiplier were not replayed. The selected frontier and all five assumptions remain unchanged.
Base:
e54523517d0f03e58352e647c403a5af3e525a57. @aliromman supplied the PR7663 construction and unchanged guide/reset source with substantial OpenAI Codex assistance. @CubeJerry supplied the selected primitive and prior typed reset in PR #7586. The source notice retains the candidate and selected predecessors' credit and licenses. Ali Romman directed this new selector, cost proof and verification with substantial OpenAI Codex assistance. New work is Apache-2.0.