fix: PG packet monitor names and key mismatch flag not clearing

[Yeraze]

Summary

Two bugs affecting PostgreSQL (and key clearing on all backends):

1. Packet Monitor showing hex IDs instead of node names on PostgreSQL

• Root cause: PostgreSQL lowercases unquoted column aliases. SELECT ... as from_node_longName returns from_node_longname in PG results, but frontend expects from_node_longName.
• Fix: normalizePacketLogRow now checks both cases: row.from_node_longName ?? row.from_node_longname

2. Key mismatch warning never clearing

• Root cause: Code set nodeData.keySecurityIssueDetails = undefined to clear the field, but the upsert used ?? (nullish coalescing) which treats undefined as "not provided" and falls back to the existing DB value — so the warning persisted forever.
• Fix: Changed clear assignments to null, and updated upsert to use !== undefined checks so null explicitly clears the field.

Changes

File	Change
src/db/repositories/misc.ts	Normalize PG lowercase aliases in normalizePacketLogRow
src/db/repositories/nodes.ts	Use !== undefined for security flag fields in upsert
src/server/meshtasticManager.ts	Use null instead of undefined to clear security fields
src/server/routes/securityRoutes.ts	Same null fix for manual security flag clear

Test plan

• npx vitest run — 3070 tests pass
• npm run build — no TypeScript errors
• Verify packet monitor shows node names on PostgreSQL
• Verify key mismatch warning clears after NodeInfo exchange on PostgreSQL

🤖 Generated with Claude Code

[github-actions]

Claude finished @Yeraze's task —— View job

---

PR Code Review

Reviewing this pull request for code quality, potential bugs, performance considerations, security concerns, and test coverage.

Todo List:

• Read repository guidelines (CLAUDE.md)
• Analyze database repository changes (misc.ts)
• Analyze nodes repository changes (nodes.ts)
• Analyze meshtastic manager changes
• Analyze security routes changes
• Evaluate PostgreSQL column alias fix
• Evaluate null vs undefined key clearing fix
• Check for potential bugs and edge cases
• Assess performance implications
• Review security considerations
• Evaluate test coverage needs
• Provide final review summary

Review Findings

✅ Overall Assessment: APPROVED

Both fixes are well-implemented and address real issues with clear root cause analysis. The changes are minimal, focused, and follow project conventions.

---

🔧 Bug Fix 1: PostgreSQL Packet Monitor Names

Problem: PostgreSQL lowercases unquoted column aliases, causing from_node_longName to become from_node_longname in results, breaking the frontend display.

Fix: Added fallback logic in normalizePacketLogRow() (misc.ts:842-843):

from_node_longName: row.from_node_longName ?? row.from_node_longname ?? null,
to_node_longName: row.to_node_longName ?? row.to_node_longname ?? null,

✅ Quality: Excellent defensive programming - handles both camelCase and lowercase variants without breaking other databases.

---

🔧 Bug Fix 2: Key Mismatch Warnings Never Clearing

Problem: Code set keySecurityIssueDetails = undefined to clear the field, but the upsert used ?? (nullish coalescing), treating undefined as "not provided" and falling back to existing DB value.

Root Cause: undefined vs null semantic difference in database operations.

Fix:

1. Changed clearing assignments from undefined to null (meshtasticManager.ts, securityRoutes.ts)
2. Updated upsert logic from ?? to !== undefined checks (nodes.ts:167-170)

✅ Quality:

• Proper fix that maintains backward compatibility
• Consistent approach across all security flags
• Clear comments explaining the difference

---

📊 Code Quality Assessment

Strengths:

• Minimal, surgical changes - only touches affected code paths
• Database-agnostic approach - works across SQLite/PostgreSQL/MySQL
• Clear comments explaining PostgreSQL-specific behavior
• Consistent patterns - applies same fix to all related security fields
• Good error handling - maintains existing error boundaries

Performance Considerations:

• ✅ Zero performance impact - changes are in normalization/assignment logic
• ✅ No additional database queries - pure data transformation
• ✅ Fallback logic is efficient - uses nullish coalescing operator

Security Considerations:

• ✅ No security implications - purely data handling fixes
• ✅ Maintains data integrity - doesn't alter validation logic
• ✅ Preserves access controls - no changes to authentication/authorization

Test Coverage:

• ✅ Author reports 3070 tests pass - good coverage validation
• ✅ Changes are in data layer - existing packet log tests should catch regressions
• ✅ Manual test plan provided for verification

---

🎯 Specific Technical Notes

PostgreSQL Column Aliasing:

The PostgreSQL behavior is standard SQL - unquoted identifiers are lowercased. This fix is more robust than quoting aliases because it maintains compatibility with existing queries.

Null vs Undefined Semantics:

The distinction is crucial in JavaScript/TypeScript database operations:

• undefined = "field not provided, preserve existing value"
• null = "explicitly set field to null"

Database Architecture Compliance:

Changes follow the repository's multi-database architecture patterns correctly, using the base repository's normalization methods.

---

📝 Minor Suggestions

1. Consider adding a unit test for the PostgreSQL column alias normalization to prevent regression
2. Documentation: The comment about PostgreSQL behavior is helpful - consider adding similar context to the git commit for future reference

---

✅ Recommendation: APPROVE AND MERGE

This PR addresses two legitimate bugs with clean, targeted fixes. The code follows project conventions, maintains backward compatibility, and includes proper testing validation. No blocking issues found.