Repository navigation
F-0b: dissolve stale XL-0N namespace admission - #9797
Merged
Merged
Conversation
briansrls
approved these changes
Aug 31, 2026
briansrls
left a comment
briansrls
left a comment
Contributor
There was a problem hiding this comment.
APPROVED — exact head c09e21d only. Verified as one commit directly atop main@6ef73334e0b1a480ae9cc4cd26bfa2ec512cbcec, changing only src/v1/stage0/src/namespace_wave_admission.rs: the fired #9719 TargetChanged admission is removed and NAMESPACE_TRANSITION_ADMISSIONS returns to the empty fail-closed roster. No semantic widening is admitted.
Merge authority becomes live only when CI for this exact head/current tested merge completes fully green and required-witnesses-floor positively passes the namespace-wave-admission phase with zero stale admissions and zero unadjudicated deltas. This review supplies the requested fresh dashboard approval. GitHub must still report this exact head and the PR mergeable at actuation. Any head change, skipped/missing namespace phase, red check, or superseding base test voids this approval.
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 31, 2026
…made before the static-debug write; merge main (stale XL-0N admission dissolved by #9797) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 31, 2026
…n that reddened this branch's floor lane); generated artifacts regenerated from the merged sources The two conflicts were both GENERATED artifacts (DESIGN.md, docs/design-ledgers.md) and the generated-artifact merge driver refused them as designed: both paths unmerged, no conflict markers in the worktree. Hand-resolving a projection would author bytes the model did not derive, so they are regenerated from the merged .dag sources instead -- main_wet, exit 0, peak RSS 8.29 GiB. The .dag sources themselves auto-merged cleanly.
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 31, 2026
…n that reddened this branch's floor lane)
This was referenced Aug 31, 2026
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Aug 31, 2026
…havior over a derived permutation population -- CloneMinimality x CopyRealizationCost axes, production source->emission path, sharded execution, calibration matrix falsifiers, findings blocking; no production repairs (#9781) * EMIT-COST-QUAL-0: grid authority, fixture/mutant generator, hermetic witnesses (draft) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * EMIT-COST-QUAL-0: wet transport, wet witnesses, inert workflow job; parse fixes Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * EMIT-COST-QUAL-0: driver as its own non-emitted module; workflow closure conjunct Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * EMIT-COST-QUAL-0: disconnect verdict typed as observation-unavailable; exact emitted-module clippy attribution Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * EMIT-COST-QUAL-0: conform to the source-annotation channel (module-scope leading blocks only) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * EMIT-COST-QUAL-0: manifest features insurance; Display-normalized runner results (drop result_is_shared) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Rebalance runner_case_block as stepwise lets (fixes 443:953 RParen parse refusal) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Exhaustive matches for the 16 NFR-flagged fns; fix Unit-inference fold, ambiguous classify_call, defaulted args Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Enroll the 12 EMIT-COST-QUAL-0 wet identities as typed route-gap debt (IsExecutable/NoMockResponse) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Fix chunk_05 row layout: no newline after tail:, drop duplicated closer Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Land the wet battery as a plain claim module (floor_changed_witness blocks inert-lane wet witnesses) The 12 wet transactions route-gap hermetically (IsExecutable/NoMockResponse) and v2.workflow.floor_changed_witness rightly blocks changed witness identities without a terminal verdict — so the battery moves out of *_test.dag discovery into gunbc.emit_copy_qualification_wet_battery, run by the inert job via explicit claim_batch --functions; converting it to floor-enrolled witnesses is part of the activation edit. Reverts the route-gap chunk_05 enrollment (identities no longer declared witnesses; enrolled-but-not-executed would red the floor). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Bind battery conversion + sole-consumer constraint into the inert standing row (parent conditions) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Attach the floor-standing annotation to the first battery fn (trailing block named no subject) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8yC * Instrument fixes from the first full battery run: pattern-aware hidden-deep-copy wrapper, pre-window clone planting, emitted lib path - hidden-deep-copy wrapper skips loop cases (subject arg is the element list, not the String) - redundant-clone mutants plant 'let input = input.clone();' BEFORE the window: a clone inside the capturing closure clones a borrowed value and is not statically redundant, so clippy was right to stay quiet; the pre-window shadowing clone kills the original unused - grounding receipt reads out/src/lib.rs (the compile door emits a full crate under src/) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Persist per-shard verdict receipts (typed refusal on write failure); qualification_case accessor Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Regenerate witnesses.yml mirror (main_wet): the inert emit-copy-qualification-battery job renders Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Instrument round 2: constant-plateau judge (+2 discriminators), AliasThenMutate linear on structural share, per-shard target dir (diagnostic-replay hazard), window-line clone planting, static-arm debug receipt Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * string_length takes its named argument Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Hoist the AliasThenMutate growth match out of the record-field position Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Round 3: linear bound 8x pricing, persistent-map linear expectation, nondeterministic behavior cause, static-arm receipt survives the sweep, clippy --all-targets Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Differential calibrations against the clean shard; hidden-deep-copy moves to the fully-qualifying sequence carrier The emitted modules THEMSELVES carry 40-50 clippy redundant-clone findings per shard (the battery's flagship static finding), so an absolute no-redundant baseline tested the emitter, not the instrument. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Steepen the quadratic growth fixture past the 128B/elem bound (crosses at n=64) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Attribution matches clippy's crate-relative file names; receipts dir made before the static-debug write; merge main (stale XL-0N admission dissolved by #9797) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Clone calibrations key on the finding's FILE (runner-planted vs emitted-module clones); refusal carries emitted_item Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * Plant redundant clones in clippy's canonical firing shape (dead original into a sink fn, no closure) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * ecq: String-typed calibration plant (redundant_clone is blind to Rc-family clones) Measured on the share shard: clean and mutant produced byte-equal finding counts (122/46) because the planted clone was carrier-typed (Rc-realized) and clippy's redundant_clone, being MIR-based, deliberately skips Rc-family clones. A minimal src/bin repro with a String clone fires (2 hits, clippy 0.1.93), so the plant is now a pre-window String-typed redundant clone — carrier-independent, runtime-arm-neutral — and the boundary is documented at the plant and in the battery comments. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * ecq: hoist calibration plant above the case loops (redundant_clone is loop-blind) Matrix repro isolated the suppressor: the identical String clone-last-use shape fires at fn top level (TOP_CASE=2) and with a custom global allocator (ALLOC_CASE=2) but is silent inside nested for-loops (LOOP_CASE=0). The prior commit's Rc-blindness attribution was a hypothesis, not a measurement — loop suppression alone explains every observation; comments and the boundary note now state only what the matrix measured. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC * ecq: byte quantities ride std.measure ByteSize (review 57849) AllocObservationRow.allocated_bytes, ConsecutivePairScan.prev_bytes, and the observed_bytes/allowed_bytes pairs on GrowthExceeds and QualificationRefusedExcessCopy are now ByteSize, compared through measure_le (byte_size_gt/byte_size_ne helpers) and summed with measure_add. The one Int -> ByteSize crossing for allocator readings is allocated_byte_size, with the Nat hop at a return position per std.checked_arithmetic. The linear bound stays Int deliberately: it is a rate (bytes/element), lifted into ByteSize only when multiplied by a cardinality delta at its single consuming site. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01JE2ZWLWkrsHvbnb8wMJ7yC --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Fable 5 <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
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.
Removes the one exact
type_reference_declaration_refTargetChanged admission after its dissolve-on trigger fired when #9719 merged. The row is now stale on every unrelated PR and blocks the namespace-wave admission phase despite zero unadjudicated deltas (observed on #9771 run 33368922338). The roster returns to its documented empty, fail-closed state; any new namespace delta remains unadjudicated until explicitly rostered.