Repository navigation
Reject unsound binary-frame verifier inputs - #101
Merged
Mister-Meeseeks merged 1 commit intoOct 9, 2026
Merged
Conversation
Bind serialized scatter gates to paid output incidences, validate physical frame domains, and refuse assertion-disabled Python runtimes. Exercise both shared checkers independently and refresh derived source hashes.
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The shared binary-frame checkers now reject words with uncharged cancelling
scatter XORs or negative frame-index aliases, and refuse Python modes that
disable their assertions. These cases could previously pass finite verification
without establishing the claimed physical word and cost.
Both replay and native profiler preparation bind scatter gates to the exact
paid output-incidence multiset. Permutations of commuting scatter gates remain
valid. The negative-test helper checks each checker separately, so the first
rejection cannot hide acceptance by the second.
Related: #64.
rfu08 identified the cancelling-scatter counterexample in its separate audit,
reviewed at
2d970ab3ca0271cfb2b609dc9aae1b7073b7774d. This submission repairsthe shared entry points used by the retained skip-frame, joint-dual and
pair-assembly checks; it does not incorporate that PR's machine-transfer work.
Validation: the pre-fix regression run failed 12 subcases. All 16 final focused
tests pass on Python 3.12.10, including cancelling/repeated/omitted scatter,
negative indices,
-O,-OO,PYTHONOPTIMIZE=1, and valid reordering.Full
make verifypasses on Ubuntu 24.04, GCC 13.3.0 and Python 3.13.12:78 isolated modules, 522 test cases and 20 independent patch checks.
Correctness and independent adversarial reviews found no issues.
A raw-byte comparison of all 1,499 baseline files confirms that circuit words,
profiles, numerical certificate fields, upstream sources, references and patches
are unchanged. Only derived provenance hashes accompany the source changes.
The unchanged Lean projects were not built. This corrects finite checkers and
adds no exponent bound or proof of the full multiplication theorem.
AI assistance: implementation and review assisted by Codex.