fix: add 13 missing reserved keywords to identifier quoting (#328)

[tianzhou]

Summary

pgschema dump was not quoting 13 PostgreSQL reserved keywords when used as column identifiers, producing invalid SQL that fails on re-apply.

The reservedWords map in ir/quote.go was missing these keywords that are classified as RESERVED_KEYWORD or TYPE_FUNC_NAME_KEYWORD in PostgreSQL's gram.y:

analyse, analyze, asc, concurrently, desc, full, leading, localtime, localtimestamp, notnull, overlaps, placing, session_user

The fix adds all 13 keywords to the map, cross-referenced against PostgreSQL's gram.y grammar file.

Fixes #328

Test plan

• Added dump test TestDumpCommand_Issue328MissingReservedKeywords with a table using all 13 keywords as column names
• Test fails before fix (columns emitted unquoted)
• Test passes after fix (columns properly quoted)
• Full dump test suite passes (no regressions)
• Run: go test -v ./cmd/dump -run TestDumpCommand_Issue328

🤖 Generated with Claude Code

[greptile-apps]

Greptile Summary

This PR fixes a critical bug where 13 PostgreSQL reserved keywords were not being quoted when used as column identifiers in pgschema dump output, resulting in invalid SQL that fails to execute.

Key Changes:

• Added 13 missing reserved keywords to reservedWords map in ir/quote.go: analyse, analyze, asc, concurrently, desc, full, leading, localtime, localtimestamp, notnull, overlaps, placing, session_user
• All keywords verified against PostgreSQL's gram.y grammar file (both reserved_keyword and type_func_name_keyword categories)
• Added comprehensive integration test with table using all 13 keywords as column names
• Added local copies of PostgreSQL's gram.y and scan.l for future reference
• Updated documentation to reflect new reference files and test counts

Impact:
The fix ensures pgschema dump output can be successfully re-applied to PostgreSQL without syntax errors when these keywords are used as identifiers.

Confidence Score: 5/5

• This PR is safe to merge with no risk
• The changes are well-tested, thoroughly researched against PostgreSQL's official grammar, and follow established patterns in the codebase. All 13 keywords were verified in internal/gram.y, the fix is minimal and focused, and comprehensive test coverage was added. The implementation is straightforward (adding entries to an existing map) with no risk of regressions.
• No files require special attention

Important Files Changed

Filename	Overview
ir/quote.go	Added 13 missing reserved keywords in alphabetical order, matching PostgreSQL's gram.y
cmd/dump/dump_integration_test.go	Added comprehensive test for issue #328 verifying all 13 keywords are quoted
testdata/dump/issue_328_missing_reserved_keywords/pgschema.sql	Expected pgschema output with all 13 keywords properly quoted

_{Last reviewed commit: b82af81}

[Copilot]

Pull request overview

Fixes invalid SQL emitted by pgschema dump when PostgreSQL reserved keywords are used as column identifiers by ensuring those identifiers are quoted consistently.

Changes:

• Add 13 missing PostgreSQL reserved keywords to ir/quote.go’s reservedWords map so they get quoted when used as identifiers.
• Add a new dump integration regression test case (testdata + TestDumpCommand_Issue328MissingReservedKeywords) covering all 13 keywords as column names.
• Update Claude/skill documentation and add local PostgreSQL lexer reference (internal/scan.l) to support future grammar/keyword investigations.

Reviewed changes

Copilot reviewed 11 out of 12 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
ir/quote.go	Extends reservedWords with the 13 missing keywords so needsQuoting triggers quoting.
cmd/dump/dump_integration_test.go	Adds an integration test to ensure dump output quotes the newly-added reserved keywords.
testdata/dump/issue_328_missing_reserved_keywords/raw.sql	New raw schema fixture using all 13 keywords as quoted column names.
testdata/dump/issue_328_missing_reserved_keywords/pgschema.sql	New expected pgschema dump output verifying quoted identifiers.
testdata/dump/issue_328_missing_reserved_keywords/pgdump.sql	New pg_dump-based setup SQL to reproduce the schema in tests.
testdata/dump/issue_328_missing_reserved_keywords/manifest.json	Metadata for the new dump regression test suite.
internal/scan.l	Adds a local copy of PostgreSQL’s lexer source as a reference artifact.
CLAUDE.md	Updates repository guidance, including references to local Postgres sources and dependencies.
.claude/skills/run_tests/SKILL.md	Updates testing guidance to reflect diff.sql/plan files naming.
.claude/skills/postgres_syntax/SKILL.md	Updates syntax-reference guidance to prefer local internal/gram.y / internal/scan.l.
.claude/skills/fix_bug/SKILL.md	Notes that --generate produces plan artifacts alongside diff.sql.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.