Exercise 2: Real Content-Type Detection Across Five Extensions — Possible Solution
====================================================================

THE TEST FILES
------------------------------
    files = {
        'app.js':    'console.log("hi");',
        'data.json': '{"a": 1}',
        'notes.txt': 'plain text notes',
        'style.css': 'body { color: blue; }',
        'page.html': '<p>hi</p>',
    }
    # each written for real to the static directory, then requested
    # over a real, live CLI-served server

VERIFIED, REAL RESULT
------------------------------
    app.js:     status=200  content-type='text/javascript'   match=True
    data.json:  status=200  content-type='application/json'  match=True
    notes.txt:  status=200  content-type='text/plain'         match=True
    style.css:  status=200  content-type='text/css'           match=True
    page.html:  status=200  content-type='text/html'          match=True

All five real files, served through the exact same serve_static()
handler, produced the correct real Content-Type header for their own
extension — with zero manual extension-to-MIME-type mapping written
anywhere in this framework's own code.

WHY THIS WORKS
------------------------------
mimetypes.guess_type() reads from a real, standard, OS-level table of
extension-to-MIME-type mappings (the same one most other software on the
machine consults) rather than anything this framework maintains itself.
Passing it the real resolved file path is enough — it inspects the
extension, not the file's actual byte content, so a genuinely correct
Content-Type falls out for any of the huge number of extensions Python's
own mimetypes module already knows about, with the framework needing
only the one guess_type() call and an application/octet-stream fallback
for anything it doesn't recognize.

WHY THIS WORKS AS AN ANSWER
------------------------------
It verifies real HTTP response headers, not just the return value of
mimetypes.guess_type() called directly — confirming the value genuinely
makes it all the way from a real file on disk, through safe_static_path()
and serve_static(), through Response.to_wsgi(), and into a real,
observable HTTP header on a live socket. Testing five genuinely different
extensions (rather than just one) confirms the mapping is real and
general, not a single lucky match.