fix(test): resolve Biome noUnsafeOptionalChaining lint errors (#1341)

[github-actions]

CI Auto-Fix

Original PR: #1341
Failed CI Run: Non-Main Branch CI/CD
Failed Job: dev-build-deploy (bun run lint step)

Root Cause

bun run lint exited 1 with 2 errors (lint/correctness/noUnsafeOptionalChaining) in tests/unit/instrumentation-crash-handler.test.ts. The Biome spec floats (@biomejs/biome: ^2, no lockfile) and resolved to 2.5.4, whose noUnsafeOptionalChaining rule flags process.report?.writeReport because the optional-chained result flows into an unconditional .mock access (would throw TypeError if process.report were undefined). The 196 useOptionalChain findings are warnings and do not fail the build.

Fix Applied

File	Line	Change	Type
tests/unit/instrumentation-crash-handler.test.ts	134, 149	process.report?.writeReport -> process.report!.writeReport	lint correctness

process.report is guaranteed present here: writeReport is only spied when process.report exists (beforeEach), and excludeEnv is asserted on the preceding line. noNonNullAssertion is off in biome.json, so the non-null assertion is permitted. Runtime behavior is unchanged.

Verification

• bun run typecheck passes (exit 0)
• bun run lint passes (exit 0, 2 errors resolved; only pre-existing warnings/infos remain)
• vitest run tests/unit/instrumentation-crash-handler.test.ts -> 9/9 tests pass
• No logic changes; only ?. -> ! on a guaranteed-present object

---

Auto-generated by Claude AI

Greptile Summary

This PR introduces a single commit on top of the perf/full-path-benchmark-optimization branch that fixes two Biome 2.5.4 lint/correctness/noUnsafeOptionalChaining errors in the crash-handler test suite. The 129 files listed in the PR diff belong to the underlying branch, not this fix commit.

• Changed lines (2): process.report?.writeReport → process.report!.writeReport in the two "genuine errors" test cases; the optional-chain form was flagged because .mock is accessed unconditionally on the chain result, which would throw TypeError if process.report were undefined.
• Correctness of the fix: In the first test, expect(process.report?.excludeEnv).toBe(true) on the preceding line already acts as an implicit non-null guard. In the second test, the beforeEach spy setup only runs when process.report exists, and Node.js always surfaces process.report. The noNonNullAssertion rule is confirmed \"off\" in biome.json.

Confidence Score: 5/5

The change is a two-line test-only substitution that replaces optional-chain syntax with non-null assertions; it introduces no logic changes and restores CI green.

The only touched file is a test file, and both non-null assertion sites are correctly backed — the first by an explicit toBe(true) guard on the preceding line, the second by the beforeEach spy condition that only runs when process.report exists. The noNonNullAssertion rule is confirmed disabled in biome.json. No source files are modified.

No files require special attention; tests/unit/instrumentation-crash-handler.test.ts is the only changed file in the fix commit and the substitution is mechanically correct.

Important Files Changed

Filename	Overview
tests/unit/instrumentation-crash-handler.test.ts	Two ?. → !. replacements on process.report.writeReport to resolve Biome noUnsafeOptionalChaining; both sites are correctly guarded (by assertion or beforeEach spy condition), and noNonNullAssertion is off in biome.json.

Sequence Diagram

%%{init: {'theme': 'neutral'}}%%
sequenceDiagram
    participant Test
    participant beforeEach
    participant processReport as process.report (Node built-in)
    participant uncaughtException as crash handler

    beforeEach->>processReport: if process.report exists → vi.spyOn writeReport
    processReport-->>beforeEach: spy installed

    Test->>uncaughtException: uncaughtException(error)
    uncaughtException->>processReport: "report.excludeEnv = true"
    uncaughtException->>processReport: report.writeReport(filename, toSafeCrashError(error))
    processReport-->>uncaughtException: empty string (mocked)

    Test->>processReport: expect(process.report?.excludeEnv).toBe(true) — implicit guard
    Test->>processReport: process.report!.writeReport — non-null assertion safe
    Test->>Test: assert reportCalls[0][1] matches expected error

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
sequenceDiagram
    participant Test
    participant beforeEach
    participant processReport as process.report (Node built-in)
    participant uncaughtException as crash handler

    beforeEach->>processReport: if process.report exists → vi.spyOn writeReport
    processReport-->>beforeEach: spy installed

    Test->>uncaughtException: uncaughtException(error)
    uncaughtException->>processReport: "report.excludeEnv = true"
    uncaughtException->>processReport: report.writeReport(filename, toSafeCrashError(error))
    processReport-->>uncaughtException: empty string (mocked)

    Test->>processReport: expect(process.report?.excludeEnv).toBe(true) — implicit guard
    Test->>processReport: process.report!.writeReport — non-null assertion safe
    Test->>Test: assert reportCalls[0][1] matches expected error

Loading

_{Reviews (1): Last reviewed commit: "fix(test): resolve unsafe optional chain..." | Re-trigger Greptile}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3d9f63954c

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Handle local DB errors before retrying hedge providers

In hedge mode, when a Drizzle query error that is not a pool-admission error is thrown while an attempt is being prepared or handled, databaseError is non-null but errorCategory remains SYSTEM_ERROR because categorizeErrorAsync only maps admission errors. The code then records an endpoint failure and falls through to provider failure/alternative launch logic, even though switching providers cannot fix a local DB outage and can poison endpoint health. Mirror the non-hedge path by overriding the category or branching on any databaseError before endpoint accounting.

Useful? React with 👍 / 👎.