For the thumb listings such as the index page, gallery page a call
to cloudfiles is made for every thumbnail on the page.

CloudFilesStorage should be changed to save container\_url on
*init*, and then use self.\_resolve\_filepath() to generate the
filepath without verifying that the file exists, just like
BasicFileStorage does.

The MIMEtype application/octet-stream is always set for files in
the public\_store.

pre word¶
=========

While reviewing the late changes, I found some things, I would like
to be considered.

edit\_media() references attachments.¶
--------------------------------------

``edit_media()`` in ``edit/view.py`` references some attachment
stuff. Is that still valid?

delete\_media\_files() might use itervalues¶
--------------------------------------------

::

    - for handle, listpath in media['media_files'].items():
    + for listpath in media['media_files'].itervalues():

import export code: Use the init/ code¶
---------------------------------------

::

    - setup_global_and_app_config(args.conf_file)
    + global_config, app_config = setup_global_and_app_config(args.conf_file)
    ...
    - config, validation_result = read_mediagoblin_config(args.conf_file)
    - connection, db = setup_connection_and_db_from_config(
    -   config['mediagoblin'], use_pymongo=True)
    + connection, db = setup_connection_and_db_from_config(app_config, use_pymongo=True)

final word¶
===========

These are UNTESTED suggestions to consider. They might break
everything, be completely wrong, or just useful. Take your one and
comment on it. Or create branch and add it here for merging.