This involves a few changes to extractor scripts and test data:
* Extractors that relied on the unintentional iteration over unrecognized
barcodes in PDF documents have been ported to the new explicit API.
* Extractor filters using the old "Barcode" type have been changed to use
other plain text or binary data mimetypes.
* Extractor types or filter types referring to ical calendar data now
separate between the whole calendar and individual events.
* Modified times in test results are updated in cases where we now have
better context times available (mainly top-level pkpass files).