DocTest Functional Tests

This file documents and tests doctest-based functional tests and basic Zope web-application functionality.

Request/Response Functional Tests

You can create Functional tests as doctests. Typically, this is done by using a script such as src/zope/app/testing/dochttp.py to convert tcpwatch recorded output to a doctest, which is then edited to provide explanation and to remove uninyeresting details. That is how this file was created.

Here we'll test some of the most basic types of access.

First, we'll test accessing a protected page without credentials:

>>> print http(r"""
... GET /@@contents.html HTTP/1.1
... """)
HTTP/1.1 401 Unauthorized
Cache-Control: no-store, no-cache, must-revalidate
Content-Length: ...
Content-Type: text/html;charset=utf-8
Expires: Mon, 26 Jul 1997 05:00:00 GMT
Pragma: no-cache
WWW-Authenticate: basic realm="Zope"
<BLANKLINE>
<!DOCTYPE html PUBLIC ...

Here we see that we got:

  • A 404 response,
  • A WWW-Authenticate header, and
  • An html body with an error message
  • Some technical headers to keep squid happy

Note that we used ellipeses to indicate ininteresting details.

Next, we'll access the same page with credentials:

>>> print http(r"""
... GET /@@contents.html HTTP/1.1
... Authorization: Basic mgr:mgrpw
... """)
HTTP/1.1 200 OK
Content-Length: ...
Content-Type: text/html;charset=utf-8
<BLANKLINE>
<!DOCTYPE html PUBLIC ...

Important note: you must use the user named "mgr" with a password "mgrpw".

And we get a normal output.

Next we'll try accessing site management. Since we used "/manage", we got redirected:

>>> print http(r"""
... GET /++etc++site/@@manage HTTP/1.1
... Authorization: Basic mgr:mgrpw
... Referer: http://localhost:8081/
... """)
HTTP/1.1 303 See Other
Content-Length: 0
Content-Type: text/plain;charset=utf-8
Location: @@contents.html
<BLANKLINE>

Note that, in this case, we got a 303 response. A 303 response is the prefered response for this sort of redirect with HTTP 1.1. If we used HTTP 1.0, we'd get a 302 response:

>>> print http(r"""
... GET /++etc++site/@@manage HTTP/1.0
... Authorization: Basic mgr:mgrpw
... Referer: http://localhost:8081/
... """)
HTTP/1.0 302 Moved Temporarily
Content-Length: 0
Content-Type: text/plain;charset=utf-8
Location: @@contents.html
<BLANKLINE>

Lets visit the page we were redirected to:

>>> print http(r"""
... GET /++etc++site/@@contents.html HTTP/1.1
... Authorization: Basic mgr:mgrpw
... Referer: http://localhost:8081/
... """)
HTTP/1.1 200 OK
Content-Length: ...
Content-Type: text/html;charset=utf-8
<BLANKLINE>
<!DOCTYPE html PUBLIC ...

Finally, lets access the default page for the site:

>>> print http(r"""
... GET / HTTP/1.1
... Authorization: Basic mgr:mgrpw
... """)
HTTP/1.1 200 OK
Content-Length: ...
Content-Type: text/html;charset=utf-8
<BLANKLINE>
<!DOCTYPE html PUBLIC ...

Access to the object system

You can use the getRootFolder() function:

>>> root = getRootFolder()
>>> root
<zope.app.folder.folder.Folder object at ...>

You can intermix HTTP requests with regular Python calls. Note, however, that making an http() call implied a transaction commit. If you want to throw away changes made in Python code, abort the transaction before the HTTP request.

>>> print http(r"""
... POST /@@contents.html HTTP/1.1
... Authorization: Basic mgr:mgrpw
... Content-Length: 73
... Content-Type: application/x-www-form-urlencoded
...
... type_name=BrowserAdd__zope.app.folder.folder.Folder&new_value=f1""")
HTTP/1.1 303 See Other
Content-Length: ...
Content-Type: text/html;charset=utf-8
Location: http://localhost/@@contents.html
<BLANKLINE>
<!DOCTYPE html ...

Now we can see that the new folder was added:

>>> list(root.keys())
[u'f1']