Skip to content

TESTING.md

Latest commit

 

History

History
165 lines (123 loc) · 3.55 KB

TESTING.md

File metadata and controls

165 lines (123 loc) · 3.55 KB
Raw

Quick Start Testing Guide

This guide will help you quickly test the GitLab AI Assistant extension.

Prerequisites

  1. ✅ Dependencies installed (npm install completed)
  2. ✅ TypeScript compiled (npm run compile completed)
  3. ✅ No lint errors

Quick Test Setup

1. Install the Extension

# Package the extension
npm run package

# This creates a .vsix file
# Install it via VS Code: Extensions → ... → Install from VSIX

2. Minimal Configuration

Press Cmd+Shift+P and run: GitLab AI: Configure Settings

Or manually set in VS Code settings:

{
  "gitlabAI.gitlabUrl": "https://gitlab.com",
  "gitlabAI.gitlabToken": "YOUR_GITLAB_TOKEN",
  "gitlabAI.llmProvider": "openai",
  "gitlabAI.llmApiKey": "YOUR_OPENAI_KEY",
  "gitlabAI.llmModel": "gpt-4"
}

3. Test in Development Mode

Instead of packaging, you can test directly:

  1. Open this folder in VS Code
  2. Press F5 to start debugging
  3. A new VS Code window will open with the extension loaded
  4. Open a Git repository in that window
  5. Test the commands

Quick Tests

Test 1: Configuration Test

Cmd+Shift+P → GitLab AI: Configure Settings → Test Configuration

Expected: ✅ GitLab and LLM connections verified

Test 2: Create Branch

Cmd+Shift+P → GitLab AI: Create New Branch
Description: "add user login feature"
Type: feature

Expected: Branch feature/XXX-add-user-login-feature created

Test 3: Commit with AI

  1. Make a small code change
  2. Cmd+Shift+P → GitLab AI: Commit Changes with AI Message
  3. Review the generated commit message Expected: Conventional commit message generated

Test 4: Review Code

  1. Make some changes (don't commit)
  2. Cmd+Shift+P → GitLab AI: Review Current Changes
  3. Check Problems panel Expected: AI review comments appear

Testing with Ollama (No API Key Needed)

If you don't want to use paid APIs:

  1. Install Ollama: https://ollama.ai
  2. Start Ollama and pull a model: 
    ollama serve
ollama pull llama2
  3. Configure: 
    {
  "gitlabAI.llmProvider": "ollama",
  "gitlabAI.llmApiUrl": "http://localhost:11434",
  "gitlabAI.llmModel": "llama2"
}

Common Issues

"Cannot find module" errors

npm install
npm run compile

"GitLab API Error"

  • Check your GitLab token has api scope
  • Verify GitLab URL includes https://

"LLM API Error"

  • Verify API key is correct
  • Check you have credits/quota
  • For Ollama, ensure ollama serve is running

Development Workflow

# 1. Make changes to src/
vim src/commands/createBranch.ts

# 2. Compile
npm run compile

# 3. Test in debug mode
# Press F5 in VS Code

# 4. Check for errors
npm run lint

Next Steps

Once basic testing works:

  1. Test the full workflow command
  2. Create a business context file
  3. Test MR creation (requires GitLab repo)
  4. Test code review with business context

Packaging for Distribution

# Create installable .vsix
npm run package

# Share the .vsix file or publish to marketplace

Debugging Tips

  • Check Output panel: View → Output → Select "GitLab AI Assistant"
  • Use Developer Tools: Help → Toggle Developer Tools
  • Enable verbose logging in extension settings
  • Test with a simple repository first

Success Criteria

✅ Extension activates without errors ✅ Configuration test passes
✅ Can create a branch with AI naming ✅ Can generate commit messages ✅ Code review generates suggestions ✅ No TypeScript compilation errors ✅ No ESLint errors (warnings OK)

For detailed documentation, see README.md