chore: combined merge of #364, #367, #365, #366, #368, #369, #370

.agents/skills/testing-fixtures/SKILL.md

@@ -144,6 +144,27 @@ This generates two tests per case (pretty and non-pretty) and uses Jest snapshot
 3. Run `npx jest` to verify all tests pass
 4. Commit the `.sql` file, `generated.json`, and any new generated test files in `__tests__/kitchen-sink/`
 
+## PostgreSQL Regression Corpus (`__fixtures__/regress/`)
+
+A ratchet over the SQL that libpg_query vendors from PostgreSQL (`test/sql/postgres_regress`,
+`test/sql/plpgsql_regress`) plus its own test SQL (`test/sql/deparse*`, `test/plpgsql_samples.sql`):
+
+- `statements.json.gz`: unique parseable statements keyed like `generated.json`
+  (`postgres_regress/arrays-12.sql`), plus the libpg_query commit they came from. Built by
+  `cd packages/deparser && npm run fixtures:regress -- /path/to/libpg_query`. Check out the
+  `x-publish.libpgQueryCommit` of the pinned `libpg-query` release first.
+- `known-failures.deparser.json` / `known-failures.plpgsql-deparser.json`: statements that do not
+  round-trip yet.
+- Tests: `packages/deparser/__tests__/regress/regress-corpus.test.ts` (one test per source file,
+  pretty and non-pretty, via `TestUtils.expectAstMatch`) and
+  `packages/plpgsql-deparser/__tests__/regress-corpus.test.ts`. They run in the normal `test` script.
+
+A new failure fails the test, and so does a known failure that now passes. After fixing deparser
+bugs (or bumping `libpg-query` and regenerating the corpus), refresh the lists with
+`UPDATE_KNOWN_FAILURES=1 npx jest __tests__/regress` (deparser) and
+`UPDATE_KNOWN_FAILURES=1 npx jest __tests__/regress-corpus` (plpgsql-deparser), then review the diff.
+Bug fixes still get a kitchen-sink fixture; the corpus only guards against regressions.
+
 ## PL/pgSQL Fixtures (`plpgsql-deparser`)
 
 The PL/pgSQL deparser has its own fixture pipeline:

__fixtures__/kitchen-sink/misc/alter-statistics.sql

@@ -0,0 +1,9 @@
+ALTER STATISTICS ab1_a_b_stats RENAME TO ab1_a_b_stats_new;
+ALTER STATISTICS "My Schema"."My Stats" RENAME TO "New Stats";
+ALTER STATISTICS public.s1 RENAME TO s2;
+ALTER STATISTICS s1 SET SCHEMA s2;
+ALTER STATISTICS s1 OWNER TO r1;
+ALTER STATISTICS ab1_a_b_stats SET STATISTICS 0;
+ALTER STATISTICS IF EXISTS ab1_a_b_stats_new SET STATISTICS -1;
+ALTER STATISTICS s1 SET STATISTICS DEFAULT;
+ALTER STATISTICS "My Schema"."My Stats" SET STATISTICS 5;

__fixtures__/kitchen-sink/misc/alter-trigger-rename.sql

@@ -0,0 +1,5 @@
+-- ALTER TRIGGER name ON table RENAME TO (RenameStmt OBJECT_TRIGGER keeps subname)
+-- Ref: PostgreSQL regress triggers.sql
+ALTER TRIGGER p ON grandp RENAME TO q;
+ALTER TRIGGER "Mixed Trig" ON "My Schema"."My Table" RENAME TO "New Trig";
+ALTER TRIGGER modified_a ON main_table RENAME TO modified_modified_a;

__fixtures__/kitchen-sink/misc/copy-option-values.sql

@@ -0,0 +1,5 @@
+-- COPY option values keep their meaning
+COPY t FROM STDIN (HEADER false);
+COPY t FROM STDIN (HEADER true, FREEZE false);
+COPY t FROM STDIN (FORMAT "CSV");
+COPY t FROM STDIN (FORMAT 'csv', HEADER off);

__fixtures__/kitchen-sink/misc/fetch-with-ties.sql

@@ -0,0 +1,10 @@
+-- FETCH FIRST n ROWS WITH TIES (SelectStmt.limitOption)
+-- Ref: PostgreSQL regress limit.sql
+SELECT thousand FROM onek WHERE thousand < 5 ORDER BY thousand FETCH FIRST 2 ROW WITH TIES;
+SELECT thousand FROM onek WHERE thousand < 5 ORDER BY thousand FETCH FIRST ROWS WITH TIES;
+SELECT thousand FROM onek WHERE thousand < 5 ORDER BY thousand OFFSET 2 ROWS FETCH NEXT (1 + 1) ROWS WITH TIES;
+CREATE VIEW limit_thousand_v_1 AS SELECT thousand FROM onek WHERE thousand < 995 ORDER BY thousand FETCH FIRST 5 ROWS WITH TIES OFFSET 10;
+CREATE VIEW limit_thousand_v_3 AS SELECT thousand FROM onek WHERE thousand < 995 ORDER BY thousand FETCH FIRST (NULL+1) ROWS WITH TIES;
+CREATE VIEW limit_thousand_v_4 AS SELECT thousand FROM onek WHERE thousand < 995 ORDER BY thousand FETCH FIRST NULL ROWS ONLY;
+SELECT * FROM t ORDER BY a FETCH FIRST $1 ROWS WITH TIES;
+SELECT * FROM t ORDER BY a FETCH FIRST -1 ROWS WITH TIES;

__fixtures__/kitchen-sink/misc/json-query-functions.sql

@@ -0,0 +1,21 @@
+SELECT JSON_EXISTS(jsonb '{"a": 1}', '$.a');
+SELECT JSON_EXISTS(jsonb '{"a": [1,2,3]}', 'strict $.a[5]' ERROR ON ERROR);
+SELECT JSON_EXISTS(jsonb '[1,2,3]', '$[*] ? (@ > $x)' PASSING 2 AS x, 'b' AS "Y" UNKNOWN ON ERROR);
+SELECT JSON_EXISTS(js FORMAT JSON, '$.a' TRUE ON ERROR) FROM t;
+SELECT JSON_VALUE(jsonb '"aaa"', '$');
+SELECT JSON_VALUE(jsonb '123', '$' RETURNING int);
+SELECT JSON_VALUE(jsonb '"2017-02-20"', '$' RETURNING date DEFAULT '2000-01-01' ON EMPTY NULL ON ERROR);
+SELECT JSON_VALUE(jsonb '[1,2]', 'strict $[*]' DEFAULT 0 + 1 ON ERROR);
+SELECT JSON_VALUE(js, '$.a' PASSING x + 1 AS x RETURNING text ERROR ON EMPTY ERROR ON ERROR) FROM t;
+SELECT JSON_QUERY(jsonb '[1,2]', '$[*]');
+SELECT JSON_QUERY(jsonb '"aaa"', '$' RETURNING text KEEP QUOTES);
+SELECT JSON_QUERY(jsonb '"aaa"', '$' RETURNING text OMIT QUOTES ON SCALAR STRING);
+SELECT JSON_QUERY(jsonb '[1,2]', '$[*]' WITH WRAPPER);
+SELECT JSON_QUERY(jsonb '[1,2]', '$[*]' WITH CONDITIONAL ARRAY WRAPPER);
+SELECT JSON_QUERY(jsonb '[1,2]', '$[*]' WITH UNCONDITIONAL WRAPPER);
+SELECT JSON_QUERY(jsonb '[1,2]', '$[*]' WITHOUT WRAPPER);
+SELECT JSON_QUERY(jsonb '[]', '$[*]' EMPTY ON EMPTY);
+SELECT JSON_QUERY(jsonb '[]', '$[*]' EMPTY ARRAY ON EMPTY EMPTY OBJECT ON ERROR);
+SELECT JSON_QUERY(js FORMAT JSON, '$.a' PASSING 1 AS x RETURNING bytea FORMAT JSON ENCODING UTF8 WITH CONDITIONAL WRAPPER KEEP QUOTES DEFAULT '"x"' ON EMPTY ERROR ON ERROR) FROM t;
+CREATE VIEW json_query_v AS SELECT JSON_QUERY(jsonb '{"a": 1}', '$.a' RETURNING jsonb WITH WRAPPER NULL ON ERROR) AS q, JSON_VALUE(jsonb '1', '$' RETURNING int) AS v, JSON_EXISTS(jsonb '1', '$') AS e;
+CREATE TABLE json_check (js jsonb CHECK (JSON_EXISTS(js, '$.a')), v text DEFAULT JSON_VALUE(jsonb '"x"', '$'));

__fixtures__/kitchen-sink/misc/json-table.sql

@@ -0,0 +1,11 @@
+SELECT * FROM JSON_TABLE(jsonb '[1, 2]', '$[*]' COLUMNS (a int PATH '$'));
+SELECT * FROM JSON_TABLE(jsonb '[1, 2]', '$[*]' COLUMNS (id FOR ORDINALITY, a int)) jt;
+SELECT * FROM JSON_TABLE(jsonb '[1, 2]', '$[*]' AS root COLUMNS (a int PATH '$')) AS jt (x);
+SELECT * FROM JSON_TABLE(jsonb '{"a": [1]}', '$' PASSING 1 AS x, 'y' AS "Y" COLUMNS (a int PATH '$.a[*] ? (@ > $x)' DEFAULT -1 ON EMPTY ERROR ON ERROR) ERROR ON ERROR) jt;
+SELECT * FROM JSON_TABLE(jsonb '{"a": 1}', '$' COLUMNS (e bool EXISTS PATH '$.a', e2 int EXISTS PATH '$.b' TRUE ON ERROR)) jt;
+SELECT * FROM JSON_TABLE(jsonb '{"a": "x"}', '$' COLUMNS (js jsonb FORMAT JSON PATH '$.a' WITH CONDITIONAL WRAPPER OMIT QUOTES, js2 text FORMAT JSON ENCODING UTF8 PATH '$' WITHOUT WRAPPER KEEP QUOTES EMPTY OBJECT ON ERROR)) jt;
+SELECT * FROM JSON_TABLE(jsonb '{"a": "x"}', '$' COLUMNS (q jsonb PATH '$.a' WITH WRAPPER EMPTY ARRAY ON EMPTY, "Mixed Col" text PATH '$.a' NULL ON ERROR)) jt;
+SELECT * FROM JSON_TABLE(jsonb '{"a": [{"b": 1}]}', '$' COLUMNS (id FOR ORDINALITY, NESTED PATH '$.a[*]' COLUMNS (b int PATH '$.b', NESTED PATH '$.c[*]' AS c_path COLUMNS (c text)), NESTED '$.d' AS d COLUMNS (d int))) jt;
+SELECT jt.* FROM t, LATERAL JSON_TABLE(t.js, '$[*]' COLUMNS (a text PATH '$.a')) AS jt;
+SELECT jt.* FROM t CROSS JOIN JSON_TABLE(t.js FORMAT JSON, '$[*]' COLUMNS (a text)) jt;
+CREATE VIEW json_table_view AS SELECT * FROM JSON_TABLE(jsonb 'null', 'lax $[*]' PASSING 1 + 2 AS a, json '"foo"' AS "b c" COLUMNS (id FOR ORDINALITY, "int" int PATH '$', js json PATH '$', jsb jsonb PATH '$', exists1 bool EXISTS PATH '$.aaa', NESTED PATH '$[1]' AS p1 COLUMNS (a1 int, NESTED PATH '$[*]' AS "p1 1" COLUMNS (a11 text)))) jt;

__fixtures__/kitchen-sink/misc/operator-names.sql

@@ -0,0 +1,18 @@
+SELECT a OPERATOR(pg_catalog.+) b FROM t;
+SELECT a OPERATOR("My Schema".+) b FROM t;
+SELECT a OPERATOR("S".===) b OPERATOR(public.*) c FROM t;
+SELECT OPERATOR("S".-) a FROM t;
+SELECT a OPERATOR("S".=) ANY (ARRAY[1]) FROM t;
+SELECT a OPERATOR("S".=) ALL (SELECT 1) FROM t;
+SELECT - -1, @ -1, 1 - -1, 2 ^ - 1;
+CREATE OPERATOR "My Schema".=== (LEFTARG = int, RIGHTARG = int, PROCEDURE = f, COMMUTATOR = OPERATOR("My Schema".===), NEGATOR = OPERATOR("My Schema".!==));
+CREATE OPERATOR === (LEFTARG = int, RIGHTARG = int, PROCEDURE = f, COMMUTATOR = ===, NEGATOR = !==);
+ALTER OPERATOR === (int, int) SET (COMMUTATOR = OPERATOR("My Schema".===));
+ALTER OPERATOR "S".=== (int, int) SET (NEGATOR = !==);
+DROP OPERATOR "S".< (int, int);
+COMMENT ON OPERATOR "My Schema".=== (int, int) IS 'x';
+COMMENT ON OPERATOR - (NONE, int) IS 'x';
+CREATE OPERATOR CLASS c FOR TYPE int USING btree AS OPERATOR 1 "S".< (int, int), OPERATOR 2 <=, FUNCTION 1 f(int, int);
+ALTER OPERATOR FAMILY f USING btree ADD OPERATOR 1 "S".< (int, int);
+ALTER TABLE t ADD EXCLUDE USING gist (c WITH OPERATOR("S".&&));
+ALTER TABLE t ADD EXCLUDE USING gist (c WITH &&);

__fixtures__/kitchen-sink/misc/show-quoting.sql

@@ -0,0 +1,11 @@
+-- SHOW keeps identifier quoting (VariableShowStmt.name)
+SHOW "TimeZone";
+SHOW "Foo"."Bar";
+SHOW "my.ext".setting;
+SHOW "ALL";
+SHOW ALL;
+SHOW TIME ZONE;
+SHOW TRANSACTION ISOLATION LEVEL;
+SHOW SESSION AUTHORIZATION;
+SHOW search_path;
+SHOW plpgsql.extra_warnings;

__fixtures__/kitchen-sink/misc/tablespace-clauses.sql

@@ -0,0 +1,19 @@
+-- TABLESPACE / USING INDEX TABLESPACE / INCLUDE / WITH on table and constraint definitions
+-- Ref: PostgreSQL regress tablespace.sql, create_table.sql, constraints.sql
+CREATE TABLE testschema.foo (i int) TABLESPACE regress_tblspace;
+CREATE TABLE t (a int) ON COMMIT DROP TABLESPACE ts;
+CREATE TABLE t (a int) USING heap TABLESPACE ts;
+CREATE TABLE testschema.dflt (a int PRIMARY KEY USING INDEX TABLESPACE regress_tblspace) PARTITION BY LIST (a) TABLESPACE pg_default;
+CREATE TABLE testschema.part_1 PARTITION OF testschema.part FOR VALUES IN (1) TABLESPACE "My Space";
+CREATE TABLE t (a int, b int, PRIMARY KEY (a) INCLUDE (b) WITH (fillfactor=10) USING INDEX TABLESPACE ts DEFERRABLE);
+CREATE TABLE t (a int UNIQUE NULLS NOT DISTINCT USING INDEX TABLESPACE "My Space");
+CREATE TABLE t (a int, CONSTRAINT u UNIQUE (a) INCLUDE (a) WITH (fillfactor=50));
+CREATE TABLE t (a int, EXCLUDE USING gist (a WITH =) INCLUDE (a) WITH (fillfactor=10) USING INDEX TABLESPACE ts WHERE (a > 0));
+ALTER TABLE testschema.atable ADD CONSTRAINT anindex_t UNIQUE (column1) USING INDEX TABLESPACE regress_tblspace;
+ALTER TABLE t ADD CONSTRAINT pk PRIMARY KEY USING INDEX "My Index";
+ALTER TABLE t ADD UNIQUE USING INDEX idx DEFERRABLE INITIALLY DEFERRED;
+CREATE TABLESPACE "My Space" OWNER regress_user LOCATION '/tmp' WITH (random_page_cost = 3, seq_page_cost = '1.5');
+ALTER TABLESPACE "My Space" SET (random_page_cost = 1.5, effective_io_concurrency = 2);
+ALTER TABLESPACE "My Space" RESET (random_page_cost);
+ALTER TABLESPACE "My Space" SET ("Random_Page_Cost" = 2);
+DROP TABLESPACE IF EXISTS "My Space";

__fixtures__/kitchen-sink/misc/transaction-chain.sql

@@ -0,0 +1,9 @@
+-- COMMIT/ROLLBACK AND CHAIN (TransactionStmt.chain)
+-- Ref: PostgreSQL regress transactions.sql
+START TRANSACTION ISOLATION LEVEL REPEATABLE READ, READ WRITE, DEFERRABLE;
+COMMIT AND CHAIN;
+ROLLBACK AND CHAIN;
+COMMIT AND NO CHAIN;
+ROLLBACK AND NO CHAIN;
+END AND CHAIN;
+ABORT AND CHAIN;

__fixtures__/kitchen-sink/misc/xmlserialize-indent.sql

@@ -0,0 +1,6 @@
+-- XMLSERIALIZE ... INDENT (XmlSerialize.indent)
+-- Ref: PostgreSQL regress xml.sql
+SELECT xmlserialize(DOCUMENT '<foo><bar><val x="y">42</val></bar></foo>' AS text INDENT);
+SELECT xmlserialize(CONTENT '<foo><bar><val x="y">42</val></bar></foo>' AS text INDENT);
+SELECT xmlserialize(DOCUMENT '<foo><bar><val x="y">42</val></bar></foo>' AS text NO INDENT);
+SELECT xmlserialize(CONTENT doc AS varchar INDENT) FROM xmltest;

__fixtures__/kitchen-sink/pretty/misc.sql

@@ -232,4 +232,9 @@ GRANT EXECUTE ON FUNCTION dashboard_private.uuid_generate_seeded_uuid TO PUBLIC;
 
 
 -- https://github.com/constructive-io/pgsql-parser/issues/217
-SELECT CAST(t.date AT TIME ZONE $$America/New_York$$ AS text)::date FROM tbl t;
+SELECT CAST(t.date AT TIME ZONE $$America/New_York$$ AS text)::date FROM tbl t;
+
+-- FETCH/MOVE ALL (FETCH_ALL is 2147483647 in the WASM AST)
+FETCH ALL FROM c;
+
+MOVE BACKWARD ALL IN c;

__fixtures__/plpgsql-generated/generated.json

@@ -165,6 +165,10 @@
   "plpgsql_deparser_fixes-65.sql": "CREATE FUNCTION test_alias_trigger() RETURNS trigger\nLANGUAGE plpgsql AS $$\nDECLARE\n  prior ALIAS FOR old;\n  updated ALIAS FOR new;\nBEGIN\n  updated.updated_at := now();\n  RETURN updated;\nEND$$",
   "plpgsql_deparser_fixes-66.sql": "CREATE FUNCTION test_assign_named_args() RETURNS text\nLANGUAGE plpgsql AS $$\nDECLARE\n  v_body text;\nBEGIN\n  v_body := ast_helpers.create_function(v_schema_name := 'app_public', v_function_name := 'f');\n  RETURN v_body;\nEND$$",
   "plpgsql_deparser_fixes-67.sql": "CREATE FUNCTION test_assign_targets() RETURNS void\nLANGUAGE plpgsql AS $$\nDECLARE\n  a int[] := ARRAY[1, 2];\n  r record;\nBEGIN\n  SELECT 1 AS f INTO r;\n  a[2] := coalesce(nullif(3, 0), 4);\n  r.f := greatest(1, 2);\nEND$$",
+  "plpgsql_deparser_fixes-68.sql": "CREATE FUNCTION test_multiline_literals(p_on boolean) RETURNS void\nLANGUAGE plpgsql AS $$\nBEGIN\n  IF p_on THEN\n    RAISE NOTICE 'line one\n  line two';\n    EXECUTE format('SELECT 1\n       FROM %I', 'pg_class');\n    PERFORM E'it''s\\\n  escaped', $q$dollar\n   quoted$q$;\n  END IF;\nEND$$",
+  "plpgsql_deparser_fixes-69.sql": "CREATE FUNCTION test_quoted_char_type() RETURNS \"char\"\nLANGUAGE plpgsql AS $$\nDECLARE\n  c \"char\" := 'a';\nBEGIN\n  RETURN c;\nEND$$",
+  "plpgsql_deparser_fixes-70.sql": "CREATE FUNCTION test_case_empty_else(x int) RETURNS text\nLANGUAGE plpgsql AS $$\nBEGIN\n  CASE x\n    WHEN 1 THEN RETURN 'one';\n    ELSE\n  END CASE;\n  RETURN 'other';\nEND$$",
+  "plpgsql_deparser_fixes-71.sql": "CREATE FUNCTION test_nested_comment_literal(p_on boolean) RETURNS void\nLANGUAGE plpgsql AS $$\nBEGIN\n  IF p_on THEN\n    PERFORM /* outer /* inner */ it's */ 'line one\n  line two';\n  END IF;\nEND$$",
   "plpgsql_control-1.sql": "do $$\nbegin\n  -- basic case\n  for i in 1..3 loop\n    raise notice '1..3: i = %', i;\n  end loop;\n  -- with BY, end matches exactly\n  for i in 1..10 by 3 loop\n    raise notice '1..10 by 3: i = %', i;\n  end loop;\n  -- with BY, end does not match\n  for i in 1..11 by 3 loop\n    raise notice '1..11 by 3: i = %', i;\n  end loop;\n  -- zero iterations\n  for i in 1..0 by 3 loop\n    raise notice '1..0 by 3: i = %', i;\n  end loop;\n  -- REVERSE\n  for i in reverse 10..0 by 3 loop\n    raise notice 'reverse 10..0 by 3: i = %', i;\n  end loop;\n  -- potential overflow\n  for i in 2147483620..2147483647 by 10 loop\n    raise notice '2147483620..2147483647 by 10: i = %', i;\n  end loop;\n  -- potential overflow, reverse direction\n  for i in reverse -2147483620..-2147483647 by 10 loop\n    raise notice 'reverse -2147483620..-2147483647 by 10: i = %', i;\n  end loop;\nend$$",
   "plpgsql_control-2.sql": "do $$\nbegin\n  for i in 1..3 by 0 loop\n    raise notice '1..3 by 0: i = %', i;\n  end loop;\nend$$",
   "plpgsql_control-3.sql": "do $$\nbegin\n  for i in 1..3 by -1 loop\n    raise notice '1..3 by -1: i = %', i;\n  end loop;\nend$$",

__fixtures__/plpgsql/plpgsql_deparser_fixes.sql

@@ -837,3 +837,48 @@ BEGIN
   a[2] := coalesce(nullif(3, 0), 4);
   r.f := greatest(1, 2);
 END$$;
+
+-- Test 68: multi-line string literals keep their exact text inside nested blocks
+CREATE FUNCTION test_multiline_literals(p_on boolean) RETURNS void
+LANGUAGE plpgsql AS $$
+BEGIN
+  IF p_on THEN
+    RAISE NOTICE 'line one
+  line two';
+    EXECUTE format('SELECT 1
+       FROM %I', 'pg_class');
+    PERFORM E'it''s\
+  escaped', $q$dollar
+   quoted$q$;
+  END IF;
+END$$;
+
+-- Test 69: quoted "char" type keeps its quotes
+CREATE FUNCTION test_quoted_char_type() RETURNS "char"
+LANGUAGE plpgsql AS $$
+DECLARE
+  c "char" := 'a';
+BEGIN
+  RETURN c;
+END$$;
+
+-- Test 70: CASE with an empty ELSE keeps the ELSE
+CREATE FUNCTION test_case_empty_else(x int) RETURNS text
+LANGUAGE plpgsql AS $$
+BEGIN
+  CASE x
+    WHEN 1 THEN RETURN 'one';
+    ELSE
+  END CASE;
+  RETURN 'other';
+END$$;
+
+-- Test 71: nested block comments do not hide the string literal after them
+CREATE FUNCTION test_nested_comment_literal(p_on boolean) RETURNS void
+LANGUAGE plpgsql AS $$
+BEGIN
+  IF p_on THEN
+    PERFORM /* outer /* inner */ it's */ 'line one
+  line two';
+  END IF;
+END$$;