Structure for testing views via webtest

[Christopher Allan Webber]

We need to start testing views to make sure they work right. Right
now we should create a "test app" to start testing things in.

Use webtest:
`http://pythonpaste.org/webtest/ <http://pythonpaste.org/webtest/>`_

The testing app should be importable from something like:

            from mediagoblin.tests.tools import get\_test\_app
            test\_app, webtest\_app = get\_test\_app()




where the test\_app is the MediaGoblinApp instance and webtest\_app
is a wrapped instance in webtest.

The testapp should do the following things:


-  create a new database with the database name
   '**mediagoblinunittests**' or something with the default connection
   settings, dumping an old database if it exists
-  Attach a mediagoblin.app.MediaGoblinApp instance to webtest with
   that database... see the structure of paste\_app\_factory to see
   how to do this
-  create a public\_store and a queue\_store that write to
   temporary directories. Use tempfile.mkdtemp() to create these
   directories
-  when calling setup\_celery\_from\_config it should make sure the
   mongodb database that it calls is the *mediagoblinunittests*
   database we created.

One note: cache the database connection in a global variable like
mediagoblin.tests.tools.DATABASE\_CONNECTION and also note whether
celery is set up in mediagoblin.tests.tools.CELERY\_IS\_SET\_UP...
we should only set up celery once :)

[Christopher Allan Webber]

Sigh, my >>> got borked.

::

    from mediagoblin.tests.tools import get_test_app
    test_app, webtest_app = get_test_app()

are the lines I meant to put in that thing that was quoted 3
times.

[Christopher Allan Webber]

Ping, any progress on this?

[Tomaz Kovacic]

Hey, I've setup the environment and passed trough the codebase. I
was planning to get started this weekend but something else came up
so it'll have to wait for a couple of days.

[Christopher Allan Webber]

Okay, thanks for the update, and let us know. This is a complicated
ticket to start with and not very well explained. Let me know if
this is an overwhelming task. We might need it soon, so if you can
get started on it over the next week that would be great.

No pressure for you to push yourself though. If the need comes up
and you can't get to it quite in time I might be able to get to it
on the week of the 16th and we can assign you to something that's
not as overwhelming for a new contributor.

All that said I don't want to discourage you from working on this
task! Your help is greatly appreciated. :)

[Tomaz Kovacic]

Thank you for you support. No pressure felt on my side. Started
hacking yesterday. I'll come back to you if get in any trouble.

[Christopher Allan Webber]

Okay! Good to hear.

[Elrond]

As per todays meetup, this issue is targetted for 0.0.3. Setting
appropiately.

[Christopher Allan Webber]

Tomaz, have you made any further progress on this?

This is high priority for 0.0.3... claiming for myself for now as
I'll work on it next.

[Elrond]

This is in progress, and chris certainly has accomplished a lot in
this area.
The 50% are just a wild guess. ;)

[Christopher Allan Webber]

This is working really well; I'd even be willing to say "it's done"
and any more changes that happen can just be incremental things.

I'll write up a braindump of how all this stuff works to the
mailing list this week.

[Will Kahn-Greene]

The original url for this bug was ​http://bugs.foocorp.net/issues/289 .
Relations:
#79: blocks, #80: blocks