** New features and improvements: **
* `category` field was added to items metadata to classify the type of
the item generated with each backend.
* The `tag` attribute added to the backends allows to mark the items
with a custom label.
* Two class methods, `has_caching` and `has_resuming`, are part now
of `Backend` class interface to notify whether a backend supports
caching and/or resuming of items.
** Backend improvements: **
* **jenkins**
- support blacklist of jobs
* **mediawiki**
- use API pages methods by default
* **phabricator**
- fetch and include projects data assigned to each task
* **redmine**
- fetch and include users data
* **remo**
- support new version of the API
* **supybot**
- parse messages written by special bots
** Bugs fixed: **
* Filepaths on merge commits were not captured on Git backend. This was
neccesary in those cases where merge commits only include data about
lines added and removed because the filepaths were not parsed and
included on the item data. (63)
* The `url` argument on the Gerrit backend was set to optional. It is
mandatory. Thus, it was set to positional on the argument parser. (60)
* Newer versions of Phabricator fixed a bug on API Conduit regarding
'constraints' parameter. The Phabricator client was modified to fix
this bug, too. (80)
* Python's `requests` library decompresses gzip-encoded responses, but
in some cases is only able to decompress some parts due to encoding
issues or to mixed contents. This problem was fixed downloading and
storing the orinal/raw data (compressed or decompressed) for furthed
processing.
* Jira backend did not return items in order, from oldest to newest. (89)
* Dates with invalid timezones were not parsed. In those cases, the
the dates will be converted usin UTC by default. (73)