Repository navigation
Conversation
…ly one disposition
The partition over the site projection was a COUNT equality — offered == routed +
declined_long + declined_fixture + declined_outside_gate + declined_cost_debt — over a
denominator that had itself already narrowed. DESIGN §5 names both halves: completeness is
an identity join, not a count equality; and a population removed before the partition is one
the partition cannot speak for.
Three changes, one seam:
1. THE UNIVERSE IS THE DECLARED POPULATION. Preparation drops modules two ways — the
exclusion substrings, and (since the 2026-08-29 gate cut) every module the gate closure
does not reach — and a witness declared in a dropped module was neither planned nor
declined. It now carries a disposition row, in the authority that already existed:
DeclinedDiscoveryExcluded { matched_substring } and DeclinedOutsideGateClosure. The
closure arm is kept distinct from DeclinedOutsideRequiredGate because they are removed by
different mechanisms, restored by different triggers, and differ by two orders of
magnitude — the closure population is the subject of the §4b rung drop "Required gate
reduced to the compiler floor", and one label over both would report it as a rounding
error on a nearby count.
2. THE CHECK IS AN IDENTITY JOIN. FloorDispositionJoinInexact reconciles the declared
identities against the rows they produced and names the offending identities in three
sets, through the SAME function the terminal ledger join uses
(reconcile_identity_population, generalized from reconcile_terminal_ledger to take
identities rather than one seam's row type). Duplicate detection moves from the planned
subset to the whole declared population: a duplicate whose first site declined used to
pass unnoticed. The four decline counters are derived from the rows instead of
accumulated beside them.
3. THE ARTIFACT CARRIES TWO AXES, NEVER FOLDED. The disposition TSV gains an outcome column
joined from the terminal ledger through claim_disposition; an identity that never ran
reads not_executed, which is a statement rather than a blank.
Evidence: the calibration pair is enrolled beside the terminal-ledger one — a population
that drops one identity and duplicates another, over which every count of the deleted form
is still exactly equal, and the join names both. Rung honesty: that suite is the local Rust
suite, removed from CI 2026-07-11, so the executing evidence on a push is the floor's own
run, whose announcement now carries declared= beside offered=.
Found on the way: gunbc.discovery_census claims twice in prose that a new
RequiredFloorDisposition arm must fail to compile in its wildcard-free matches.
DeclinedOutsideRequiredGate had already been added with neither match acquiring an arm and
nothing refused — its witness sits outside the gate closure, so no executing path typechecks
it. The arms are added and the claim is restated at its honest rung with its trigger.
Not built, and named as this join's next-rung triggers rather than improvised: the semantic
producer axis (no authority maps a witness to its producer) and the Rust #[ignore] roster (a
different universe with no roster authority). → docs/plans/witness-execution-closure.md
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WmatNuCFnqTdE2KoiwEcm4
# Conflicts: # src/v1/stage0/src/bin/claim_executor.rs # src/v1/stage0/src/cli_run.rs
…ssion/nimble-ibex-902
…ons to match it The CI auto-heal bot resolved the #9685 conflict by reimplementing the declared-population enumeration on the producer itself — folded over preparation's FULL module index, with the prepared closure and exclusion map classifying the returned identities. That is a better §3 answer than the split this branch carried (producer for the prepared subject, naming-hygiene scan for the removed sources), so it is kept and mine is dropped rather than restored. What it left describing something the code no longer does, corrected here: - the receipt line still cited BarrenTestSidecar, a refusal the producer replaced, and said 'offered' where the guarantee is now over every DECLARED entry; - the site loop's own comment still said the population is 'the identities preparation dropped', which is the deleted design; - the plan doc's item 1 described preparation emitting the rows. And one consequence neither the bot nor this branch had stated: folding the producer over the full index means its per-file refusals — misplaced test decl, barren sidecar, misplaced wire contract, malformed live_tree_disposition — now stop the REQUIRED floor for any module under the source roots. That is a real widening of this lane's subject. It is survivable today (measured: zero misplaced decls, zero barren sidecars tree-wide) and the first violation authored anywhere will red this lane rather than the one owning the file, so it is written into the doc rather than left for that run to discover. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WmatNuCFnqTdE2KoiwEcm4
…ssion/nimble-ibex-902
…-growth disposition (review 57430) TWO FINDINGS, BOTH REAL, ANSWERED IN THE TREE RATHER THAN IN THE PR THREAD. RETENTION. Folding the discovery authority over the full module index holds every source's bytes, and the out-of-closure majority is held by nothing else — the prepared graph is the gate closure. Leaving that on PreparedRepository restored corpus-scale retention across the longest phase of the program, in the one lane RequiredFloorGrowthBudgetStanding records as having no measured memory margin, where growth owes a named payment. No payment is claimed and none is needed: the inventory is now std::mem::take'n into the fold's own local and dropped with module_for_path the moment the rows are classified. What survives is paths, module names and function names — never bytes. SEED GROWTH. Measured at item grain, this change adds no hand Rust declaration: one function is renamed (reconcile_terminal_ledger -> reconcile_identity_population, generalized so one join serves two seams) and everything else is fields, coproduct arms and bodies of existing items. Its disposition in gunbc.seed_growth_admission's vocabulary is ExistingSeedItemModified, and the doc supplies that arm's payload: dag authority v2.workflow.required_floor RequiredFloorDisposition (the two new arms land in the model, the seed realizes them), so capability origin is ModeledCapability, not a capability originated in Rust. WHAT THE REVIEW ASKED FOR AND THIS DELIBERATELY DOES NOT SUPPLY: authored before/after Rust census figures. seed_growth_forward_freeze_policy_note records that hand-item and hand-LOC deltas are functions of the diff, that an authored copy is a second representation of a fact the diff already owns, and that a prior receipt authored exactly such figures and got them wrong. The deriving instrument (gunbc.rust_item_host_observation) and the adjudicating join (seed_growth_admit_change, which no required phase invokes yet) are named instead. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WmatNuCFnqTdE2KoiwEcm4
…e one section The bot and this branch fixed the retention finding concurrently and identically in effect (take the full inventory out of the prepared repository, drop it when the fold's rows are classified). Theirs is kept, with the views CONSUMED into the fold's rows rather than cloned and the duplicated drop/count pair removed. Its hand-Rust receipt and mine were two paragraphs about one fact in one document, which is the duplication this doc's own subject condemns. They are folded into a single section: mine supplied the disposition vocabulary (ExistingSeedItemModified, ModeledCapability), why no SeedGrowthJustification row is owed, and why authored census deltas are refused; theirs supplied the payload halves I had left out — the modified item list, the owning lane v1-hand-queue-drain, and a concrete deletion trigger. Both are required by seed_growth_forward_freeze_policy_note; neither alone answered it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WmatNuCFnqTdE2KoiwEcm4
briansrls
marked this pull request as ready for review
August 29, 2026 23:02
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
Contributor
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b8199b0aae
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Comment on lines
+3013
to
+3014
| drop(module_for_path); | ||
| drop(discovery_sources); |
There was a problem hiding this comment.
With the default GUNBC_EVAL_MEMO enabled, calls made through producer_frame memoize full argument values, including each source's content; therefore dropping only discovery_sources does not release the newly added full-tree corpus. Because producer_frame remains in scope until run_required_floor returns, those bytes stay resident throughout the long claim-execution fold, undermining the memory-margin motivation of this change and potentially pushing the required CI lane into reclaim/OOM. Drop or drain producer_frame after parse_floor_discovery_producer_result once files owns the required strings.
Useful? React with 👍 / 👎.
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.
Auto-opened by session-dashboard for session
nimble-ibex-902.Pushing to
session/nimble-ibex-902advances this PR.Worker attestation
Before flipping this PR to ready for review, confirm each item:
npm test,cargo test) and the result.Closes #Ndirective.Summary
TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.
Test plan