Redirect and url_for🔖

After saving a task, you need to send the user somewhere. The obvious move is render_template("display.html", ...). Don't do that.

The refresh problem🔖

If you render_template after a POST:

1. User submits form → POST → data saved → page renders
2. User presses browser refresh
3. Browser says "resend the POST request?"
4. User clicks OK → data saved again — a duplicate row in your CSV

This keeps happening every time they refresh. In a task tracker, that's annoying. In an e-commerce checkout, it's a billing disaster.

POST-redirect-GET🔖

The fix is to redirect after every successful POST:

@app.route("/input", methods=["GET", "POST"])
def input_task():
    if request.method == "POST":
        # ... read form, write to CSV ...
        return redirect(url_for("display"))  # <-- redirect, not render_template
 
    return render_template("input.html", projects=projects)

Now the sequence is:

1. User submits form → POST → data saved → redirect to /display
2. Browser follows the redirect → GET request to /display → page renders
3. User presses refresh → re-runs the GET, not the POST → no duplicate

This pattern is called POST-redirect-GET and it's standard practice for any form submission that modifies data.

url_for() vs a hardcoded string🔖

You could write:

return redirect("/display")

That works. But url_for("display") is better:

return redirect(url_for("display"))

url_for takes a function name (the Python function for the route, as a string) and generates the correct URL. If you ever rename the route path from /display to /tasks/all, the hardcoded string breaks silently. url_for("display") still works — it looks up the function, not the path.

The datetime stamp🔖

One last piece before you write the POST handler. Every row needs a timestamp:

from datetime import datetime
 
datetime.now().strftime("%Y-%m-%d %H:%M:%S")
# → "2025-10-07 09:15:00"

Call this inside the POST block, not at module level — you want the time of submission, not the time the server started.

The import line already includes datetime in your starter code. You just need to call it when building the row dict.