ci: eliminate redundant compile pass in test:coverage script

[github-actions]

Summary

Removes a redundant TypeScript compilation + ESLint pass from the test:coverage script, which runs on every Linux CI job. This reduces unnecessary work without changing test behavior.

Optimizations

1. Eliminate Double Compilation in test:coverage

Type: Resource / Efficiency
Impact: Saves ~3–8 seconds per Linux CI run (one tsc + one eslint src execution removed)
Risk: Low

Root Cause:

The previous test:coverage script chained two commands that each triggered compilation:

test:unit  →  npm run compile  (compile #1)  →  mocha
npm test   →  pretest: compile (compile #2) + lint  →  vscode-test

Changes:

Before:

"test:coverage": "npm run test:unit && npm test"

After:

"test:coverage": "npm run compile && npm run lint && c8 --config .c8rc.json mocha out/unit/unit.test.js && vscode-test"

The flattened script runs the full pipeline once: compile → lint → c8 mocha → vscode-test. The vscode-test binary is called directly (bypassing the pretest hook) since compilation and linting have already run.

Rationale: npm test always triggers the pretest lifecycle hook (compile && lint). Chaining test:unit (which also compiles) and then npm test caused tsc and eslint to each run twice. The flattened script produces identical results with one fewer compile+lint cycle.

Detailed Analysis

Previous execution path for npm run test:coverage on Linux CI:

1. npm run test:unit → npm run compile (compile chore(main): release 0.1.0 #1) → c8 mocha out/unit/unit.test.js
2. npm test → pretest (compile ci: remove redundant pull request branch specification #2 + lint chore(main): release 0.1.0 #1) → vscode-test

New execution path:

1. npm run compile (compile chore(main): release 0.1.0 #1)
2. npm run lint (lint chore(main): release 0.1.0 #1)
3. c8 --config .c8rc.json mocha out/unit/unit.test.js
4. vscode-test

The test:unit script is unchanged and still self-contained with its own compile step for standalone use. Only the test:coverage aggregation script is updated.

Expected Impact

• Time Savings: ~3–8 seconds per Linux CI run
• Risk Level: Low — same compile output, same tests, same coverage collection
• Standalone npm run test:unit: Unaffected, still self-contained

Testing Recommendations

• Review workflow syntax
• Confirm vscode-test binary resolves correctly in node_modules/.bin
• Monitor first few runs after merge to verify coverage report is identical

Generated by CI Optimization Coach

To install this agentic workflow, run

gh aw add githubnext/agentics/workflows/ci-coach.md@442992eda2ccb11ee75a39c019ec6d38ae5a84a2

• expires on Mar 5, 2026, 1:15 PM UTC

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.17%. Comparing base (6a1a90c) to head (8147238).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #179   +/-   ##
=======================================
  Coverage   59.17%   59.17%           
=======================================
  Files           6        6           
  Lines        1984     1984           
  Branches      143      143           
=======================================
  Hits         1174     1174           
  Misses        810      810           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Copilot]

Pull request overview

This PR optimizes the VS Code extension’s CI test pipeline by removing redundant TypeScript compilation and ESLint passes from the Linux test:coverage run, reducing CI time without changing the set of tests executed.

Changes:

• Replaces test:coverage’s chained test:unit && npm test with a flattened pipeline (compile → lint → c8+mocha → vscode-test) to avoid duplicate compile/lint execution.