Copilot Prompting: Iteration

The Iterative Process with Copilot Chat

Copilot Chat remembers the context of your current conversation. Use this to your advantage:

1. Initial Prompt: Ask Copilot Chat for what you need, providing context using #selection, #file, @workspace, etc.
2. Analyze Response: Evaluate the generated code or explanation. Is it accurate? Does it meet all constraints (like library usage, language version)? Is it efficient?
3. Identify Gaps/Errors: Pinpoint what needs improvement.
4. Refine Prompt (Follow-up): In the *same chat thread*, ask a follow-up question or give a modifying instruction. Refer implicitly ("that function", "the previous suggestion") or explicitly ("refactor the code you just gave me to...") to the previous response. Examples:
   • "Can you add input validation to that function?"
   • "Rewrite the previous example using async/await instead."
   • "That's close, but it needs to handle the case where the input array is empty."
   • "Explain the time complexity of the solution you provided."
   • "Could you show me an alternative approach using [different library/pattern]?"
5. Repeat: Continue the dialogue until the result is satisfactory.

Example Iteration Flow (Copilot Chat)

Generate a simple Express route for GET /api/items that returns a static list of item names.

app.get('/api/items', (req, res) => {
  const items = ["item1", "item2", "item3"];
  res.json(items);
});

Modify that route handler to be an async function and include a basic try/catch block that sends a 500 status on error.

app.get('/api/items', async (req, res) => {
  try {
    const items = ["item1", "item2", "item3"];
    // Simulate async operation if needed
    await new Promise(resolve => setTimeout(resolve, 50));
    res.json(items);
  } catch (error) {
    console.error("Error fetching items:", error);
    res.status(500).send("Internal Server Error");
  }
});