Events inserted with Python 2 contain unicode u'' literals

[squeaky-pl]

recurrence column of event stores either a Python serialized string or a list because it uses repr to serialize it and also ast.literal_eval to deserialize it.

SELECT e.recurrence FROM recurringevent r JOIN event e ON e.id = r.id WHERE e.recurrence LIKE BINARY '%u\'%' LIMIT 20;

recurrence
[u'RRULE:FREQ=WEEKLY;UNTIL=20160801T165959Z;BYDAY=TU']
[u'RRULE:FREQ=WEEKLY;UNTIL=20161025T150000Z;BYDAY=TU']
[u'RRULE:FREQ=WEEKLY;UNTIL=20161130T075959Z;BYDAY=MO,WE,FR']
[u'RRULE:FREQ=WEEKLY;UNTIL=20161202T075959Z;BYDAY=MO,WE,FR']
[u'RRULE:FREQ=WEEKLY;UNTIL=20160912T065959Z;BYDAY=MO']
[u'RRULE:FREQ=WEEKLY;UNTIL=20160919T035959Z;BYDAY=MO']
[u'RRULE:FREQ=WEEKLY;UNTIL=20161031T162959Z;BYDAY=MO']
[u'RRULE:FREQ=WEEKLY;UNTIL=20161212T045959Z;BYDAY=MO']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170118T045959Z;INTERVAL=2;BYDAY=WE']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170427T170000Z;BYDAY=MO,TH']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170414T204500Z;BYDAY=MO,WE,FR']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170405T223000Z;BYDAY=WE']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170419T201500Z;INTERVAL=2;BYDAY=WE']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170424T203000Z;BYDAY=MO']
[u'EXDATE;TZID=America/Los_Angeles:20161226T130000', u'RRULE:FREQ=WEEKLY;UNTIL=20170508T200000Z;BYDAY=MO']
[u'RRULE:FREQ=DAILY;UNTIL=20170629T130000Z']
[u'EXDATE;TZID=America/Los_Angeles:20161221T160000,20170104T160000', u'RRULE:FREQ=WEEKLY;UNTIL=20170206T075959Z;BYDAY=MO,WE']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170510T065959Z;BYDAY=MO,WE']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170821T065959Z;BYDAY=MO,WE']
[u'RRULE:FREQ=WEEKLY;UNTIL=20170911T065959Z;BYDAY=MO,WE']

At present with Python 3 the parser still accepts u'' literals:

>>> ast.literal_eval("[u'RRULE:FREQ=WEEKLY;UNTIL=20170118T045959Z;INTERVAL=2;BYDAY=WE']")
['RRULE:FREQ=WEEKLY;UNTIL=20170118T045959Z;INTERVAL=2;BYDAY=WE']

but it can change in the future, which would break sync engine.

• Remove u prefixes from data saved in the db
• Consider switching this field to JSON for better future compatibility