------------------
New Features:
- The .base.Exporter class now provides an own `.convert` method
which accepts `binary`, `async` and `stream` arguments
and calls the appropriate conversion method internally.
Currently, it will return
- `None`, if a `stream` was given,
- binary data, if `binary` was given (and not `stream`), and
- the document_id, if `async` was given, and
- a `dict` with all meta information contained otherwise.
Improvements:
- The .base.Exporter class now provides (and uses) a `conversionSource`
method which in turn uses the existing `converted_url` method.
**Note:** if providing text instead of a URL as the ``document`` config key,
you likely need to specify a ``baseUrl`` as well; otherwise the reactor
might fail to load stylesheets or images during conversion.
Development hints:
- When developing and testing your PDFreactor_ conversions,
be sure you have more than one Zope worker thread ready;
otherwise you might wonder about HTTP errors
when the reactor sends requests to your instance!
- We provide the original API (from the PDFreactor_ distribution) as `.raw_api`
module (for development and testing;
helps to check whether a certain problem
is related to our pythonization measures).
Just in case.
[tobiasherp]