FEAT-064: mechanize the qualified obligation-identity invariance (ObligationId.v) - #119
Open
avrabe wants to merge 1 commit into
Open
FEAT-064: mechanize the qualified obligation-identity invariance (ObligationId.v)#119avrabe wants to merge 1 commit into
avrabe wants to merge 1 commit into
Conversation
…igationId.v) proofs/rocq/ObligationId.v — admit-free, //proofs/rocq:obligationid_test PASSES. Tree-wide: 0 Admitted/admit/Axiom across 20 .v files. Proves BOTH sides, deliberately: - ordinal_stable_under_foreign_insertion (+ _anywhere) — THE CORRECTED AC: an insertion of other-kind operators leaves a later site's intra-region ordinal, hence its identity, unchanged. - ordinal_shifts_under_same_kind_insertion — WHY the qualification is needed: one same-kind insertion strictly increases the ordinal, so FEAT-064's original unqualified AC is false. - survivor_inherits_deleted_identity / aliasing_is_real — the ALIASING HAZARD as a theorem rather than a caveat: delete the first of two same-kind sites and the survivor's ordinal becomes exactly the deleted site's. This is the formal justification for FEAT-065's conservatism rule (degrade to `uncertain`, never over-claim `discharged`), so the limitation cannot quietly evaporate the next time someone reads the AC optimistically. AC#3 CORRECTED to match the evidence rather than the other way round: it claimed "id equality implies same structural site", which is a HASH property requiring collision resistance. The identity is a SHA-256 and no such claim is made — the theorems are about the KEY tuple. Scope stated in the file header: the region PATH component is not modelled (under the qualified hypothesis the region structure is untouched); its list algebra stays γ-swept, the same split Segment.v uses. Env note for the next Rocq build: nix-build was missing from PATH — the recurring macOS-rewrites-the-shell-profile issue. The store (22,871 entries) and daemon were intact; the sourcing block survives in /etc/bashrc but is gone from /etc/zshrc. Workaround used: source nix-daemon.sh per command. Durable fix needs sudo. rivet validate PASS. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
📐 rivet artifact deltaPR: #119 Base SHA: Validationhead — `rivet validate` resultbase — `rivet validate` result (for comparison)Artifact stats
full stats — headDiff (base → head)AADL model — headPosted by the |
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
1 participant
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.
Adds
proofs/rocq/ObligationId.v— admit-free,//proofs/rocq:obligationid_testpasses locally (0Admitted/admit/Axiomacross all 20.vfiles).It proves both sides, on purpose
ordinal_stable_under_foreign_insertion(+_anywhere)ordinal_shifts_under_same_kind_insertionsurvivor_inherits_deleted_identity,aliasing_is_realThat third row is the point. DD-020 recorded aliasing as prose; as a proven proposition it can't quietly evaporate the next time someone reads the AC optimistically — and it's the formal justification for FEAT-065's conservatism rule (degrade to
uncertain, never over-claimdischarged).Claim corrected to match the evidence
AC#3 previously read "
id equality implies same structural siteis proven admit-free". That is a hash property requiring collision resistance — the identity is a SHA-256 and the theorems are about the key tuple, not the digest. The AC now says what the file actually proves. Changing the claim rather than stretching the proof.Honest scope (in the file header)
The region path component isn't modelled: under the qualified hypothesis the insertion opens no region, so the structure is untouched and the path is trivially preserved. Its list algebra stays γ-swept — the same split
Segment.vuses.rivet validatePASS; core tests green.🤖 Generated with Claude Code