Fixes many issues

[Hydrocharged]

Fixes:

• Generated column with COALESCE: every INSERT fails after ALTER TABLE ADD PRIMARY KEY #3323
• Brackets are dropped from saved expressions: generated columns, defaults and CHECK constraints compute wrong results #3324
• Casting a character(n) value to text keeps its trailing spaces, so comparisons and CHECK constraints fail #3325
• convert_from() is not found #3326
• A role with SELECT on a table is refused COUNT(*) on it: "permission denied for routine count" #3327
• generation_expression in information_schema.columns is NULL for a generated column #3328
• information_schema.triggers is empty while the trigger exists and fires #3330
• A named NOT NULL column constraint refuses the whole table #3332
• A CHECK constraint that calls regexp_like with a cast argument refuses every row #3333
• Casts to regnamespace fail with "unable to resolve type" #3334
• A trigger with WHEN (old.* IS DISTINCT FROM new.*) makes every UPDATE of its table fail #3336

[github-actions]

	Main	PR
Total	42090	42090
Successful	19854	19885
Failures	22236	22205
Partial Successes1	5440	5438

	Main	PR
Successful	47.1703%	47.2440%
Failures	52.8297%	52.7560%

${\color{lightgreen}Progressions (32)}$

aggregates

QUERY: insert into bytea_test_table values(decode('ff','hex'));

QUERY: insert into bytea_test_table values(decode('aa','hex'));

alter_table

QUERY: /* End test case for bug #16242 */
/* Test case for bug #17409 */
create table attbl (p1 int constraint pk_attbl primary key);

QUERY: create table atref (c1 int references attbl(p1));

QUERY: drop table attbl, atref;

QUERY: create table attbl (p1 int constraint pk_attbl primary key);

QUERY: create table atref (c1 int references attbl(p1));

QUERY: drop table attbl, atref;

constraints

QUERY: SELECT * FROM DEFAULTEXPR_TBL;

functional_deps

QUERY: CREATE TEMP TABLE articles (
    id int CONSTRAINT articles_pkey PRIMARY KEY,
    keywords text,
    title text UNIQUE NOT NULL,
    body text UNIQUE,
    created date
);

QUERY: SELECT id, keywords, title, body, created
FROM articles
GROUP BY id;

QUERY: SELECT a.id, a.keywords, a.title, a.body, a.created
FROM articles AS a JOIN articles_in_category AS aic ON a.id = aic.article_id
WHERE aic.category_id in (14,62,70,53,138)
GROUP BY a.id;

QUERY: SELECT aic.changed
FROM articles AS a JOIN articles_in_category AS aic ON a.id = aic.article_id
WHERE aic.category_id in (14,62,70,53,138)
GROUP BY aic.category_id, aic.article_id;

regproc

QUERY: SELECT to_regnamespace('pg_catalog');

QUERY: SELECT to_regnamespace('"pg_catalog"');

QUERY: SELECT to_regnamespace('Nonexistent');

QUERY: SELECT to_regnamespace('"Nonexistent"');

strings

QUERY: insert into toasttest values(decode(repeat('1234567890',10000),'escape'));

QUERY: insert into toasttest values(decode(repeat('1234567890',10000),'escape'));

QUERY: insert into toasttest values(decode(repeat('1234567890',10000),'escape'));

QUERY: insert into toasttest values(decode(repeat('1234567890',10000),'escape'));

QUERY: SELECT decode('1234567890abcdef00', 'hex');

QUERY: SELECT decode(encode(('\x' || repeat('1234567890abcdef0001', 7))::bytea,
                     'base64'), 'base64');

QUERY: SELECT decode(encode('\x1234567890abcdef00', 'escape'), 'escape');

subselect

QUERY: select count(*) from tenk1 t
where (exists(select 1 from tenk1 k where k.unique1 = t.unique2) or ten < 0);

triggers

QUERY: SELECT trigger_name, event_manipulation, event_object_schema, event_object_table,
       action_order, action_condition, action_orientation, action_timing,
       action_reference_old_table, action_reference_new_table
  FROM information_schema.triggers
  WHERE event_object_table in ('pkeys', 'fkeys', 'fkeys2')
  ORDER BY trigger_name COLLATE "C", 2;

QUERY: UPDATE trigger_test SET f3 = 'bar';

QUERY: UPDATE trigger_test SET f3 = NULL;

QUERY: UPDATE trigger_test SET f3 = NULL;

QUERY: UPDATE trigger_test SET f3 = 'bar';

QUERY: UPDATE trigger_test SET f3 = NULL;

QUERY: UPDATE trigger_test SET f3 = NULL;

Footnotes

1. These are tests that we're marking as Successful, however they do not match the expected output in some way. This is due to small differences, such as different wording on the error messages, or the column names being incorrect while the data itself is correct. ↩

[coffeegoddd]

@Hydrocharged DOLT

read_tests	from_latency	to_latency	percent_change
covering_index_scan_postgres	2.43	2.43	0.0
groupby_scan_postgres	78.6	78.6	0.0
index_join_postgres	2.26	2.22	-1.77
index_join_scan_postgres	1.61	1.61	0.0
index_scan_postgres	467.3	475.79	1.82
oltp_point_select	0.36	0.37	2.78
oltp_read_only	6.43	6.43	0.0
select_random_points	0.72	0.72	0.0
select_random_ranges	1.04	1.04	0.0
table_scan_postgres	458.96	475.79	3.67
types_table_scan_postgres	1170.65	1149.76	-1.78

write_tests	from_latency	to_latency	percent_change
oltp_delete_insert_postgres	6.67	6.67	0.0
oltp_insert	3.36	3.36	0.0
oltp_read_write	13.46	13.46	0.0
oltp_update_index	3.55	3.55	0.0
oltp_update_non_index	3.25	3.25	0.0
oltp_write_only	7.04	7.04	0.0
types_delete_insert_postgres	7.17	7.17	0.0

[itoqa]

Commit: f35c4aa: 20 test cases ran, 1 failed ❌, 19 passed ✅.

Summary

The run covers database behavior across normal value conversions, null handling, schema and routine permissions, trigger metadata, and boundary cases such as malformed input and unusual numeric values. It exercises both expected application flows and adversarial/error paths, with broad behavior appearing sound aside from an encoding edge case that can alter invalid input silently.

Merge with caution — this PR introduces a medium-severity data-integrity issue in which malformed bytes can be accepted as replacement text instead of rejected, allowing callers to receive changed content without an error. The remaining reviewed findings are not merge blockers, but the attributable conversion failure warrants correction before treating the change as fully safe.

Tests run by Ito

Result	Severity	Type	Description
❌		Convert	Invalid bytes for BIG5, EUC_JP, and EUC_KR returned replacement characters. These inputs should be rejected with an encoding error, as invalid UTF-8 and unsupported EUC_CN inputs already are.
✅	—	General	Verified acceptable by independent adversarial review: the observed behavior is intended and documented in this codebase. Review notes: The finding invents a contract that negative int16/int32 inputs must be rejected, but the repository's explicit OID regression tests require those inputs to wrap to 4294967295 and reserve range errors for int64. Regnamespace is an OID type, and its PR-added cast implementations consistently mirror the existing OID-family behavior, so the reported result is intentional rather than a defect. The beh…
✅	—	General	The reported round-trip error came from sending returned binary bytes through a normal text parameter, not from the namespace send or receive code. A protocol-correct retry could not run because the local application container was unavailable.
✅	—	General	A low-privilege caller could use the built-in length function, but could not run an ungranted function or procedure, including through the search path.
✅	—	General	The built-in length function returned the expected result, while every ungranted user-defined function and procedure call was denied.
✅	—	General	Verified acceptable by independent adversarial review: the observed behavior is intended and documented in this codebase. Review notes: The cited return-nil branch is intentional deferred validation, not a reachable privilege bypass. GetSchemaName does not fail when an explicit schema is supplied; it fails here only when an unqualified call has no existing current schema, a state from which user-defined routines are not resolved. Built-ins remain resolvable through pg_catalog by design, and procedures independently fail during lat…
✅	—	Binding	A stored procedure bound a record containing a text value and NULL, then returned both values correctly as scalar.
✅	—	Char	Padded CHAR values compare as equal, and converting them to text, name, or varchar removes the trailing spaces.
✅	—	Decode	SQL decoding accepted hex, base64, and escape input, and converting the decoded bytes returned the text Hello.
✅	—	Decode	Malformed hex, base64, and escape input returned clear SQL errors instead of a value or a server crash.
✅	—	Distinct	Matching rows with NULL in the same position are treated as equal. The distinct check returned false, and the inverse check returned true.
✅	—	Distinct	Rows with one NULL value versus a number, and rows with different numbers, were correctly marked as different. The inverse checks also returned false.
✅	—	Distinct	Comparing rows with different numbers of values returned a clear SQL error. The database still accepted the next query.
✅	—	Expression	The query plan shows parentheses around the negative value and the nested arithmetic expression, so the expressions keep the intended order and meaning.
✅	—	Namespace	The existing qa_namespace schema resolved to a namespace value, and both the text value and an array containing it returned the expected schema name.
✅	—	Namespace	Verified acceptable by independent adversarial review: the observed behavior is intended and documented in this codebase. Review notes: The finding invents a contract that negative int2/int4 inputs must be rejected, but the repository's explicit OID cast tests require those values to wrap to 4294967295 and separately require int8 negatives to fail. Regnamespace is implemented as an OID-backed type and the new casts consistently copy the established per-source-width behavior, so the reported result is intended and should not be tre…
✅	—	Routine	The built-in version check succeeded, while a user-created routine without execute permission was rejected.
✅	—	Routine	A user-defined routine was denied before permission was granted, then returned its result after the caller received EXECUTE permission.
✅	—	Triggers	The database returned three trigger rows with the expected event, timing, order, function call, and row-level settings.
✅	—	Triggers	The trigger catalog keeps rows for INSERT, UPDATE, and DELETE events and leaves unsupported events out.

Tip

Reply with @itoqa to send us feedback on this test run.

[zachmu]

Have your agent break these into independent PRs, one per issue