@@ -70,6 +70,10 @@ At T1.12b (checkpoint 4, subagent P2-5) the task rows above were capped at about
 
 **T1.25b, its task row as it stood at T1.6b.** Evidence: Started on lane `rust-core-phase-1-lane-1a` at b455b94, rebased onto T1.24b part 1's landing; `BASE_SHA` and `CHECKPOINT_SHA` de247a3; candidate complete, both approve at acbbf11 (round 2); lands by its landing commit, whose SHA and CI run the next package records (E25). Notes: Added by the owner 2026-10-11 (E66(h)). `sdm-engine::time::text` reads text by `date-rules.md`'s "The core's text formats" (ISO 8601 and RFC 3339 and the client's two shapes, parsed by `jiff`'s `strtime`, the client's order and two-digit year rules), everything else refused by name (`text_time`; an open order `text_time_order`); inference scores at fewest and at most; DT-7's capability and `DATES_READ_DAY_FIRST` kept; contract unchanged (E68). `text.rs` 884 to 624 lines. Vectors: `text-times.json` from the spec apart from the reader (2,360 columns), `text-times-differential.json` the reference against the core over T1.25's 13,456 columns, classified. Register: DT-6 rejected, DT-15 approved by the owner's T1.25b answer, DT-16 (in the list, read otherwise) and DT-17 (inference's text outside the shapes) approved by the owner 2026-10-11 (relayed at fix round 1), no contract case affected. Measured locally: `gate.sh --base b455b94 --run python-core,python-old,rust-host` passes in `ref-pandas3` with no transition (protected paths changed: `ci.yml`, `protected-paths.txt`, `register.yaml`, the client's `test_the_schema_route.py`, `text-times.json`, `text-times-differential.json`, `text_times.rs`, `text_times_record.py`, `tests/test_contract_ci_gate.py`); the family maps and the AdventureWorks journey unchanged on both hosts. E49 probes: the reader equals the spec's Python reading on all 13,456 old columns; six seeded faults each fail the vectors; 60 generated DuckDB sources old against core (24 the reference's, 10 DT-12, 13 DT-15, 13 DT-16); no new reader of outside JSON. Carried: T1.18's floor-cell waiver removed with a control (E49 bug fix); T1.18's row says E66(f) answered T1.15's cut. Fix round 1: an open order is `text_time_order` in inference too, and class `a` needs a value outside the list (41 inference decisions to DT-16); the year written held to 1 to 9999 (see [Fix-round histories](#fix-round-histories))
 
+**T1.6b.** Fix round 1 (review of 18cdfaf: Codex changes required, one P2; the subagent's review still running): `s0.validate` quieted `DATE_LITERAL_DAY_FIRST` for every query, but the shadow check whose old validator gives the reference's warning runs only for a query the core accepts, so a refused query showed no day-first warning on the core route where the old route shows one. The run now holds a quieted warning for the caller (`run`'s `kept`, `errors.warn_from_core`'s `held`) and `s0.validate` shows the held ones when the core refuses (`errors.warn_held`); nothing is validated again. `test_a_day_first_literal_warns_on_the_core_as_on_the_old_route` holds an accepted query and two refused after the literal is read (an unknown table; a second literal naming no date); PQL-3 says which count differs. The subagent approved with P3s, the cheap ones taken: the current year is read only for a literal that places a two-digit year, so a zone the run cannot resolve refuses only such a query (P3-1; the host's `/etc/localtime` name, unchecked against its roots, then reaches only that refusal); the shared comparison fails a doubled core warning (P3-2); E69's evidence names fed8174 (P3-3).
+
+**T1.6b, its task row as it stood at T1.28w.** Evidence: Started on lane `rust-core-phase-1-lane-1a` at 87f1700, rebased onto T1.24b part 2's landing and then T1.27 part 1's; `BASE_SHA` and `CHECKPOINT_SHA` fed8174; candidate complete, for review; lands by its landing commit, whose SHA and CI run the next package records (E25). Notes: The D1 to D9 reader complete (D8, D4's warning by the platform's padding, the year in the local zone), `DATE_LITERAL_DAY_FIRST` on `validate_query`'s warnings channel, worded by the Python host and held back there only for an accepted query, whose shadow check gives the reference's own (fix round 1); `KNOWN_DIFFERING` empty; the 29 quirks, the custom-profile and validator-option rows, `dates.json`, a bounded fuzz test and a depth-999 subprocess test; DT-9 rejected; the D8 and DT-9 S0 cases in acceptance as twins, baselined; PQL-3 and PQL-4 proposed; the crate's rename held back (E69(e): the gate runs the checkpoint's L0 guards, whose data name the crate; T1.28w part 1 makes them read the names from the checked tree, part 2 renames); T1.25b's carries (E68's column, the open-order test, inference's `text_time` before an open order); E69, E70 (the owner: no local `--full`), E71 (T1.28w split). Measured locally: `gate.sh --base fed81745 --run python-core,python-old,rust-host` in `ref-pandas3` passes with the protected paths named (observations, ratchet, compare with no differences, l0); protected paths changed (25): the three baselines, the two twins, `protected-paths.txt`, the reference file, three refusal maps' pointers into `materialize.rs`, `register.yaml`, `test_the_routing_seam.py`, `test_the_rust_host.py`, `custom-profiles.json`, `known-differing.json`, `quirks.json`, `record_quirks.py`, `pql_corpus.rs`, `text-times-differential.json`, `text_times.rs`, `text_times_differential.py`, `s0_cases.rs`, and `tests/test_contract_cases.py`, `test_contract_ratchet.py` (the known-differing tests start from a pinned case) and `test_contract_register.py` (the removed-ID fault names PQL-2). E49 probes: the differential over the corpus and 330 generated queries differs only by PQL-4 (7); width, depth and long-run probes answered by name; seeded faults (D4, the zone year, a quirk, a D8 offset, the open-order check, the deep test's stack, the S0 quiet) each fail a test; as built in [as-built.md](as-built.md#t16b)
+
 ## Design task rows before T1.18's rewrite
 
 At T1.18 (checkpoint 5) these rows of design.md's task table were rewritten to state their current scope and exits, without the record of which package amended them; each is below as it stood.
 
@@ -40,8 +40,8 @@ A task's checks are active from the moment it lands, as listed in the design's t
 | T1.17 Clean-container wheel smoke | 1a | not started | | The E16 floor and newest lanes, both gating. Follow-up from T1.8's review: `docs/get-started/installation.md` tells Databricks users to pin the runtime's NumPy, which conflicts with `numpy>=1.26` on DBR 14.3 LTS (NumPy 1.23); the install guidance must be revalidated on DBR 14.3 under E16 before the epic merges |
 | T1.18 L2 in CI | 1a | done | Landed at b455b94; candidate run 38196008570 green at 7571d3e; landing runs 38197231678 (lane) and 38197231684 (phase) green; `BASE_SHA` and `CHECKPOINT_SHA` c52f2800 (3 review rounds, 2 fix rounds) | `contract-python` per runtime cell, `rust-test` once, both through the checkpoint's `gate.sh`; `contract/tests` in `ref-pandas3` on the effective expectations; H1's cell rule in `compare.py`; the floor cell report-only for T1.18's own run alone (E65(c)), gating from its landing; HOST-3 proposed; E65; ran before T1.15 and T1.17 by the orchestrator's order, T1.15's cut since answered by the owner (E66(f)); as built in [as-built.md](as-built.md#t118), notes in [Fix-round histories](#fix-round-histories) |
 | T1.19 Local gate | 1a | not started | | Shrunk by the owner (E53(b)): `gate.sh --only` and `--locked`, each locked check shown passable and failing (H26), and a seeded fault only for a check that still lacks one, the existing comparator fault tests counting; the route-proof faults named and kept: a `python-core` observation without its `HostRun` operations, a `python-old` observation with `_native` loaded, and the request-UUID control; its exit no longer requires `--locked 1a`, which T1.20 runs |
-| T1.6b PQL parser completion | 1a | in progress | Started on lane `rust-core-phase-1-lane-1a` at 87f1700, rebased onto T1.24b part 2's landing and then T1.27 part 1's; `BASE_SHA` and `CHECKPOINT_SHA` fed8174; candidate complete, for review; lands by its landing commit, whose SHA and CI run the next package records (E25) | The D1 to D9 reader complete (D8, D4's warning by the platform's padding, the year in the local zone), `DATE_LITERAL_DAY_FIRST` on `validate_query`'s warnings channel, worded by the Python host and held back there only for an accepted query, whose shadow check gives the reference's own (fix round 1); `KNOWN_DIFFERING` empty; the 29 quirks, the custom-profile and validator-option rows, `dates.json`, a bounded fuzz test and a depth-999 subprocess test; DT-9 rejected; the D8 and DT-9 S0 cases in acceptance as twins, baselined; PQL-3 and PQL-4 proposed; the crate's rename held back (E69(e): the gate runs the checkpoint's L0 guards, whose data name the crate; T1.28w part 1 makes them read the names from the checked tree, part 2 renames); T1.25b's carries (E68's column, the open-order test, inference's `text_time` before an open order); E69, E70 (the owner: no local `--full`), E71 (T1.28w split). Measured locally: `gate.sh --base fed81745 --run python-core,python-old,rust-host` in `ref-pandas3` passes with the protected paths named (observations, ratchet, compare with no differences, l0); protected paths changed (25): the three baselines, the two twins, `protected-paths.txt`, the reference file, three refusal maps' pointers into `materialize.rs`, `register.yaml`, `test_the_routing_seam.py`, `test_the_rust_host.py`, `custom-profiles.json`, `known-differing.json`, `quirks.json`, `record_quirks.py`, `pql_corpus.rs`, `text-times-differential.json`, `text_times.rs`, `text_times_differential.py`, `s0_cases.rs`, and `tests/test_contract_cases.py`, `test_contract_ratchet.py` (the known-differing tests start from a pinned case) and `test_contract_register.py` (the removed-ID fault names PQL-2). E49 probes: the differential over the corpus and 330 generated queries differs only by PQL-4 (7); width, depth and long-run probes answered by name; seeded faults (D4, the zone year, a quirk, a D8 offset, the open-order check, the deep test's stack, the S0 quiet) each fail a test; as built in [as-built.md](as-built.md#t16b) |
-| T1.28w Wire structures | 1a | not started | | Split from T1.28a after checkpoint 7 by the orchestrator (E71), on lane 1a right after T1.6b: typed wire structures in `sdm-wire`, the payload schema derived from the canonical typed definitions (checkpoint 5's contract-lock bound), dtype and config serialization, and the decision to drop a held sampler state moved into the core, said explicitly (`_rust/run.py:164-176` parses the error's field path today); no ledger rows move; then lane 1a runs T1.17, T1.15 (`--check` only), T1.19 and T1.20 |
+| T1.6b PQL parser completion | 1a | done | Landed at 1471563; candidate run 38282074831 green at 7bda7f2; landing runs 38284050562 (lane) and 38284050361 (phase) green; `BASE_SHA` and `CHECKPOINT_SHA` fed8174 (2 review rounds, 1 fix round) | The D1 to D9 reader complete (D8, D4's warning by the platform's padding, the year in the local zone), `DATE_LITERAL_DAY_FIRST` on `validate_query`'s warnings channel, worded by the Python host; `KNOWN_DIFFERING` empty; the 29 quirks, the custom-profile and validator-option rows, a bounded fuzz test and a depth-999 subprocess test; DT-9 rejected; the D8 and DT-9 S0 cases in acceptance as twins; PQL-3 and PQL-4 proposed; the crate's rename held back to T1.28w (E69(e)); E69 to E71; as built in [as-built.md](as-built.md#t16b), notes in [Fix-round histories](#fix-round-histories) |
+| T1.28w Wire structures | 1a | in progress | Started on lane `rust-core-phase-1-lane-1a` at 1471563; `BASE_SHA` and `CHECKPOINT_SHA` 1471563; in two landings: part 1 first, the L0 guards' bug fix (E56(a)), `l0_deps.py` and `l0_scan.py` reading the parser crate's name from the checked tree's own copies, so the gate's checkpoint copy can judge a rename; part 2 the rename and the wire structures; lands by its landing commit, whose SHA and CI run the next package records (E25) | Split from T1.28a after checkpoint 7 by the orchestrator (E71), on lane 1a right after T1.6b: typed wire structures in `sdm-wire`, the payload schema derived from the canonical typed definitions (checkpoint 5's contract-lock bound), dtype and config serialization, and the decision to drop a held sampler state moved into the core, said explicitly (`_rust/run.py:164-176` parses the error's field path today); no ledger rows move; then lane 1a runs T1.17, T1.15 (`--check` only), T1.19 and T1.20 |
 | T1.20 Phase 1a review | 1a | not started | | Completed independent review, blocking findings resolved; every Phase 1a check active; held-out cases written with their SHA-256 committed (H11); the S0 differential run once (H11); `python-ref` regenerated; owner check-in after it (H12); carried from T1.18's review to this check-in: a run that is not a push skips the gate, and `land.py` takes the newest run of any event (both before T1.18; `land.py` takes bug fixes only); by H1, a regression the same on both routes is reported, not gated, in a cell other than the reference's |
 | T1.21 DT-3 and ORD-1 findings | 1b | done | Landed at 18fe2e4; candidate run 38063738921 green at 28d9715; landing runs 38065017832 (lane) and 38065017590 (phase) green; `BASE_SHA` and `CHECKPOINT_SHA` fb61c00; both approve at 28d9715 (0 fix rounds) | DT-3 and ORD-1 answered by `contract/tests/test_register_probes.py`, ORD-1's stable order proposed for the owner; as built in [as-built.md](as-built.md#t121), notes in [Fix-round histories](#fix-round-histories) |
 | T1.21b AdventureWorks fixture and refusal map | 1b | done | Landed at 684669c; candidate run 38074711733 green at be933a3; landing runs 38075855462 (lane) and 38075855476 (phase) green; `BASE_SHA` and `CHECKPOINT_SHA` 18fe2e4; both approve at be933a3 (2 fix rounds; 1 same-way failure) | The AdventureWorks fixture and case, diagnostic, and its refusal map of 13 entries over 15 probes (E51); as built in [as-built.md](as-built.md#t121b), notes in [Fix-round histories](#fix-round-histories) |
@@ -216,11 +216,10 @@ The histories of landed packages moved verbatim to [fix-round-history.md](fix-ro
 - T1.18: "T1.18, its task row as it stood at T1.25b." in [fix-round-history.md](fix-round-history.md)
 - T1.26b: "T1.26b." in [fix-round-history.md](fix-round-history.md)
 - T1.25b: "T1.25b.", "T1.25b, its task row as it stood at T1.6b." in [fix-round-history.md](fix-round-history.md)
+- T1.6b: "T1.6b.", "T1.6b, its task row as it stood at T1.28w." in [fix-round-history.md](fix-round-history.md)
 
 **T1.24b.** E49 probes before the first push: (1) the route changes (every edit of a core-built schema is the core's `edit_schema`; `schema_edited` the core's), so an old-against-core differential over 102 generated DuckDB sources (52 and 50, two seeds, the scripts kept outside the repository as T1.26b's were), each with 2 to 7 edits (keys set, cleared and re-set as a name or a one-column tuple, time columns set and cleared, meanings as text and as a `Stype`, columns left out, relationships replaced, appended and popped, a missing column, a non-joinable link, a time column on a non-time type), then `materialize` and `predict` against the scripted endpoint, on the reference, the old route and the core: all 574 steps (218 refused) alike in outcome and public schema; `materialize` and `predict` the corrected reference's (P1B-13) but five `keyless_neighbors` (P1B-3) and one DT-12; one difference found and fixed (a key named on a link's own foreign key, which the reference's engine refuses; `check_links`), one registered (an unhashable value naming a column, P1B-16) and one new refusal registered (an edit after a change past the description, P1B-17); (2) the edit payload's readers (`edit_schema`'s three attachments, `Materialize.held`): 200, 5,000 and 100,000 levels and 100,000-wide lists and objects, each refused by name or read (`sdm-core`'s `tests/edit_schema.rs`); (3) the new comparators fail closed, seeded: the edit pin's check on both hosts (missing, empty, retyped, differing) and the core-agreement rule on a perturbed `edit_schema` input (by hand, `compare.py` naming `core.inputs`); a setter that skips the core fails the client's count of `edit_schema` runs; a core that took an invalid meaning fails `sdm-schema`'s tests (by hand). Rebased onto T1.18's landing (b455b94), then onto part 1's landing (de247a3, the freeze's reading fix, split out by the orchestrator so the gate could read this package's E66 allowances); the T1.26b history moved to [fix-round-history.md](fix-round-history.md).
 
-**T1.6b.** Fix round 1 (review of 18cdfaf: Codex changes required, one P2; the subagent's review still running): `s0.validate` quieted `DATE_LITERAL_DAY_FIRST` for every query, but the shadow check whose old validator gives the reference's warning runs only for a query the core accepts, so a refused query showed no day-first warning on the core route where the old route shows one. The run now holds a quieted warning for the caller (`run`'s `kept`, `errors.warn_from_core`'s `held`) and `s0.validate` shows the held ones when the core refuses (`errors.warn_held`); nothing is validated again. `test_a_day_first_literal_warns_on_the_core_as_on_the_old_route` holds an accepted query and two refused after the literal is read (an unknown table; a second literal naming no date); PQL-3 says which count differs. The subagent approved with P3s, the cheap ones taken: the current year is read only for a literal that places a two-digit year, so a zone the run cannot resolve refuses only such a query (P3-1; the host's `/etc/localtime` name, unchecked against its roots, then reaches only that refusal); the shared comparison fails a doubled core warning (P3-2); E69's evidence names fed8174 (P3-3).
-
 ## Size-trend reviews
 
 The design's "Harness scope" asks the checkpoint reviews for the size trends; the status page was cut (E53(a)), so they are recorded here and nowhere else.