fix(pg/parser): BYT-9315 — parenthesized multi-join in pg_get_viewdef output [stack 1/3]#105
Merged
rebelice merged 2 commits intoApr 22, 2026
Conversation
…bquery
Root cause: parseParenTableRef's 1-token peek couldn't distinguish
'((SELECT ...))' nested subquery from '((a JOIN b))' nested joined_table.
Code routed all '((' to subquery, breaking pg_get_viewdef() output like
'SELECT * FROM ((a JOIN b ON TRUE) JOIN c ON TRUE)' and cascading
through catalog.Exec → SQL review dead for the whole instance.
Three changes:
1. parenBeginsSubquery (T3 snapshot scan) — walks the matched paren
block, classifies outer as subquery vs joined_table based on
depth-1 content + set-op / JOIN continuation after nested close.
2. parseJoinedTable (T6/T7 hybrid nonterminal) — separates parsing
joined_table inside '(' ... ')' from generic parseTableRef,
enforces JoinExpr root so '(a)' / '((a))' / '((SELECT 1) x)'
reject per PG grammar.
3. parseSelectWithParens (T5 left-factoring) — removed
`cur == '(' { recurse }` short-circuit. Always delegates to
parseSelectNoParens so parseSelectClause's set-op precedence
climbing handles '((SELECT 1) UNION (SELECT 2))' correctly.
Effect on pg/pgregress: +17 PG official regress statements now pass
(0 new failures). FROM-clause edge-case parity with PG 17 across all
'(' dispatch shapes audited.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…seline)
Lands the starmap artifacts that will drive Phase 1-5:
- docs/plans/2026-04-21-pg-paren-dispatch.md — global plan (Codex-approved, 2 rounds)
- pg/parser/PAREN_AUDIT.{md,json} — 94 dispatch sites classified
- pg/parser/SCENARIOS-pg-paren-dispatch.md — 228 scenarios, 6 phases
Phase 0 (this commit's predecessor) is marked [x] in SCENARIOS.
Phases 1-5 execute via pg-paren-dispatch-driver / -worker skills.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Merged
5 tasks
rebelice
added a commit
that referenced
this pull request
Apr 22, 2026
… CI fence (#106) Stacked on #105. Adds PG 17 testcontainer oracle harness, 188 hand-curated probes, fuzz harness, PAREN_AUDIT (94 sites), CI workflow, audit lint. See PR for full description. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
rebelice
added a commit
that referenced
this pull request
Apr 22, 2026
…leanup (#107) Stacked on #105 #106. Closes oracle-discovered KB-1/KB-2/KB-3, audits + fixes 27+ Class A silent-accept sites, lands G1-GATE lint enforcement. See PR for full description. 🤖 Generated with [Claude Code](https://claude.com/claude-code)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes BYT-9315: a customer's view had a
pg_get_viewdef()output of shapeSELECT * FROM ((a JOIN b ON TRUE) JOIN c ON TRUE)which omni parser rejected. The failure cascaded throughcatalog.Execand killed SQL review for every statement against the instance (originally surfaced as BYT-9308).Root cause:
parseParenTableRefused a 1-token peek that couldn't distinguish((SELECT ...))(nested subquery) from((a JOIN b))(nested joined_table) — both start with((.Three structural fixes in
pg/parser/select.go:parenBeginsSubquery(T3 snapshot scan) — walks the matched paren block, classifies outer as subquery vs joined_table by depth-1 content + set-op / JOIN continuation after nested close.parseJoinedTable(T6/T7 hybrid) — separates "parse joined_table inside'(' ... ')'" from genericparseTableRef, enforces JoinExpr root.parseSelectWithParens(T5 left-factoring) — removedcur == '(' { recurse }short-circuit, letsparseSelectClause's set-op precedence handle nested paren operands.Net effect: +17 PG official regress statements pass (0 new failures). FROM-clause
(parity with PG 17 across all dispatch shapes audited.This is the foundation PR. Followed by:
Test plan
🤖 Generated with Claude Code