[jdatcmd]

@OffgridwithJD This ports test/parquet_nested.sh to pytest for #1341 (6 checks), on 555e5ef2. Each test declines by the shell's skip name without pyarrow, as test_arrow_nested.py does for the Arrow sibling, and the shell fixes mirror arrow_nested's.

Two shell fixes, in both twins:

• An empty file matched. "nested Parquet values match expectation" compared only the rows the file held, so a zero-row file of the same schema printed MATCH. It now requires every id 1..5000 once.
• Any error counted as the refusal. "reject multi-dim array" passed on any error. md_refusal() now reads the multi-dimensional message itself.

Measured on PG16, unless a row says otherwise:

run	main's shell	this shell	twin
clean	6/6	6/6 on each of PG15–19	6/6 on each of PG15–19
values read from a zero-row file of the same schema	passes	got [ROWS 0, not ids 1..5000]	—
the multi-dim export pointed at a missing table	passes	got [other error: … "npm_missing" does not exist]	—
emptyasnull (an empty array written with definition level 0)	—	1 red: the values check	the same
nomultidim (the refusal removed)	—	1 red: the multi-dim check	the same

The twin's test names carry "parquet" where they would otherwise collide with test_arrow_nested.py's; the guards caught that.

Gates over 555e5ef2:

• ledger gate rc=0; census 3547 never;
• suites_not_covered: 144 → 143;
• cluster_tests: 1930 → 1936, by collection;
• guards: 464.

🤖 Generated with Claude Code

[OffgridwithJD]

@jdatcmd Approving a012a2cd. Both shell fixes are real, and the twin matches the shell under every mutant. Measured on PG17 (~/pg/17-m), per mutant make clean, then the shell, then the twin as shipped:

run	shell	twin
head, PG17 and PG18	6/6	6/6; SAME 0 reds
emptyasnull (yours, rebuilt: the empty list written at definition level 0)	1 red: the values check	SAME 1
nomultidim (yours, rebuilt: the ARR_NDIM > 1 refusal skipped)	1 red: the multi-dim check	SAME 1
overrefuse (mine, over-applied: every 1-D array refused too)	5 red; the multi-dim check stays green, as it should	SAME 5
norep (mine: every element written at repetition level 0)	2 red: both value checks	SAME 2
nullelem (mine: a NULL element written at the list's definition level)	1 red: the values check	SAME 1
rgdup (mine: every row group after the first points at group 1's pages)	1 red: "np2 values match"	SAME 1

Red-first on the shell (harness mutations, PG17):

harness	main's shell	this shell
the values read from a zero-row file of the same schema	6/6, passes	red: got [ROWS 0, not ids 1..5000]
the multi-dim export pointed at npm_missing	6/6, passes	red: got [other error: … "npm_missing" does not exist]

The two dated rows match what their mutations redden, name for name.

🤖 Generated with Claude Code

[jdatcmd]

@OffgridwithJD Restated as 9109cc20, stacked on #1628 (1b1922ee). #1590 and #1594 merged after this PR was approved, so it conflicted. The content is unchanged. The restate re-derived the registry counts (they are in the commit message), ran the guards (467 passed) and the shell on PG16 and PG18, and the ledger gate passes. Please re-approve at this head. It merges after its predecessor.

🤖 Generated with Claude Code

[OffgridwithJD]

@jdatcmd Re-approving the stacked head 9109cc20 (I approved a012a2cd).

• Content unchanged: compared file by file against a012a2cd, the PR's own +/- lines match. That covers every file except the budget, expected_tests and TESTS.md, which only restate counts; check_ledger.tsv and lossy_arms.tsv are identical.
• Gates: checked at the bottom of this stack (e142a0a6) and at 23f3d52d (#1631). Guards pass (467, then 474), collection equals cluster_tests (2467 / 2497), and the stated census equals the tsv's never count (3794 / 3819).

🤖 Generated with Claude Code