Add regression tests for bindvar type inference with #>> operator

[fulghum]

Fixes: #3012

Depends on: dolthub/go-mysql-server#3663

[github-actions]

	Main	PR
covering_index_scan_postgres	2668.92/s	2607.36/s	-2.4%
groupby_scan_postgres	162.47/s	157.89/s	-2.9%
index_join_postgres	723.09/s	718.17/s	-0.7%
index_join_scan_postgres	930.75/s	924.05/s	-0.8%
index_scan_postgres	40.76/s	40.41/s	-0.9%
oltp_delete_insert_postgres	432.46/s	${\color{lightgreen}515.78/s}$	${\color{lightgreen}+19.2\%}$
oltp_insert	456.39/s	${\color{red}332.81/s}$	${\color{red}-27.1\%}$
oltp_point_select	4891.01/s	4859.50/s	-0.7%
oltp_read_only	4157.23/s	4271.56/s	+2.7%
oltp_read_write	2151.42/s	${\color{red}1743.65/s}$	${\color{red}-19.0\%}$
oltp_update_index	540.19/s	${\color{red}326.93/s}$	${\color{red}-39.5\%}$
oltp_update_non_index	465.13/s	${\color{red}328.95/s}$	${\color{red}-29.3\%}$
oltp_write_only	1008.61/s	${\color{lightgreen}1583.20/s}$	${\color{lightgreen}+56.9\%}$
select_random_points	2209.69/s	2267.92/s	+2.6%
select_random_ranges	1755.48/s	1825.52/s	+3.9%
table_scan_postgres	40.60/s	41.04/s	+1.0%
types_delete_insert_postgres	526.93/s	${\color{red}399.35/s}$	${\color{red}-24.3\%}$
types_table_scan_postgres	17.89/s	19.09/s	+6.7%

[itoqa]

Commit: 1e1ceb0: 16 test cases ran, 16 passed ✅.

Summary

Coverage spans database-backed JSON filtering and value extraction, including matching and nonmatching text, nested and missing fields, null handling, scalar paths, and non-text values. It also exercises parameter type inference and round-trip behavior across inserts, reads, updates, conflict updates, mixed values, and neighboring operations, with both normal flows and edge cases behaving correctly.

Safe to merge — the exercised application behaviors are healthy, with no regressions, new failures, or previously flagged failures attributable to this PR. No merge-blocking risk was identified.

Tests run by Ito

Result	Severity	Type	Description
✅	—	Operator	Comparing the stored name with a different value returned one result row with count 0, so the filter did not report a false match.
✅	—	Operator	The nested profile.name value was read as Alice, and the matching row count was 1 without query or protocol errors.
✅	—	Operator	The filter matched Alice once and returned zero matches for a missing path, JSON null, and SQL NULL.
✅	—	Operator	Shallow, nested, empty, and missing paths returned the expected matches without errors or false matches.
✅	—	Operator	The Alice value matched once. Bob, numbers, booleans, objects, arrays, and NULL values did not match.
✅	—	Path	The parameterized JSONB lookup ran successfully and found the row with name Alice.
✅	—	Path	A missing JSON path correctly returned zero matches when compared with Alice. The query completed successfully without treating the missing value as text.
✅	—	Path	A JSON null value did not match the text value Alice, and the query returned zero matches without an error.
✅	—	Path	The JSON field filter returned one matching row with count 1 and no execution or result-decoding error. The recorded check used the text wire format, so it does not demonstrate the separate binary-format path described in the test plan.
✅	—	Path	The JSONB filter accepted an unspecified parameter and returned the expected count of 1 without an execution error.
✅	—	Path	The filtered JSON record was found and the database returned one row with count 1 through the client connection.
✅	—	Planner	Insert and SELECT queries kept the right parameter types when the client sent an unspecified type. The expected rows were stored and read back without conversion errors.
✅	—	Planner	Updating an existing row and changing a conflicting insert both kept the expected value 42 and the SQL NULL flag.
✅	—	Planner	The insert accepted unspecified, explicitly typed, and NULL values together. The saved numbers, decimal value, and NULL value were all returned as expected.
✅	—	Planner	The database kept mixed parameter values in the correct positions, handled a short type list, and matched the expected JSON value. The cleanup warning did not affect the successful results.
✅	—	Planner	The dependency revision passed JSONB filtering and the neighboring INSERT, UPDATE, conflict-update, mixed-parameter, short-metadata, and NULL-handling checks.

Tip

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

[zachmu]

LGTM!

[github-actions]

	Main	PR
Total	42090	42090
Successful	18862	18863
Failures	23228	23227
Partial Successes1	5583	5583

	Main	PR
Successful	44.8135%	44.8159%
Failures	55.1865%	55.1841%

${\color{lightgreen}Progressions (1)}$

copyselect

QUERY: drop table test3;

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. ↩