Repository navigation
Conversation
…ed record as provenance encoded-body-depth-borrowing, high-cube-b1-fit-bridge, ordinary-one-factor-stock, sealed-strict-root and unconditional-batched-crt pinned certificates/selected-result.json as a live input and failed on main with working source drift. Mark that row as provenance, refresh the pins of the physical-ordinary-interchange files changed by the previous commit, and declare the two files the repaired source.py now reads. high-cube-b1-fit-bridge checked the live record bytes and kappa directly; it now checks an exact pinned copy (same blob b1fea0c) and verifies only input rows against the working tree. encoded-body-depth-borrowing's RESULT.json counts its authenticated inputs, 20 -> 21. No scientific code or result changes. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…hat contains them The runner resolves each SOURCES.json row at base_commit unless it carries an explicit pin. Rows refreshed after CrocSwap#5434 now carry head:32890bc5de034ab8fc93ac88ead7279a6a416d04:<path>, so their blobs resolve exactly. scripts/research_package.py --check-only passes for all five packages. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Requested by Mark
Change and mathematical scope
Replay repair only. No scientific change and no new κ. Follow-up to #5434. Five packages that consume
research/physical-ordinary-interchangepinnedcertificates/selected-result.jsonas a live input, so their replays fail on main withworking source drift. #5434's hosted CI re-ran them and showed the failure. The packages are:For each package:
SOURCES.jsonrow forcertificates/selected-result.jsonchanges frominputtoprovenance; the blob and sha256 stay the same. The record is also dropped from the liveVERIFY.jsoninputs.physical-ordinary-interchangefiles changed by Replay physical-ordinary-interchange against its pinned selected record (no scientific change) #5434 are refreshed. In practice that issource.py, plusSOURCES.jsonandVERIFY.jsonwhere a package pinned them.source.pyis consumed, the two files it now reads are declared as inputs:inputs/selected-result-pinned.jsonand that package'sVERIFY.json.Two packages need slightly more:
verify.pycompared the live record bytes and κ with constants. It now checks an exact pinned copy (selected-result-pinned.json, blobb1fea0c9…) with the same constants, and checks onlyinputrows against the working tree.RESULT.jsonfieldsource_inputscounts authenticated input rows; it changes from 20 to 21. Every other field is unchanged.Task and overlap
This is maintenance in the line of #5410/#5417 (jwr404) and #5434, on base
32890bc5d. No other open PR touches these packages.Validation
Locally on Windows/Python 3.12, using a POSIX-relative-path shim that is not part of the PR:
-Omode.RESULT.json.The other three packages are slow on this laptop. sealed-strict-root's science suite runs for more than 11 minutes here, while #5434 showed the CI runner about 3× faster. Hosted CI is the decisive check for them.
check_research_commit.pypasses.Credit
Package authors: Hiroki Kumazaki; Tekalign; Nick Petalas; Cinna Davis; and the high-cube-b1-fit-bridge contributors. jwr404 identified the drift. Mark (Matarener) with substantial Claude Code (Claude) assistance; Apache-2.0.
Open next questions
None justified.
🤖 Generated with Claude Code