Scrapy-poet

-------------------

* :ref:`Additional requests <additional-requests>`, when mapped to
:class:`scrapy.Request <scrapy.http.Request>` objects, now get their
``dont_filter`` parameter set to ``True``, to ask downloader middlewares like
:class:`~scrapy.downloadermiddlewares.offsite.OffsiteMiddleware` not to drop
those requests.

-------------------

* :func:`scrapy_poet.utils.testing.make_crawler` now respects setting
priorities when it receives a :class:`~scrapy.settings.Settings` object
instead of a :class:`dict`.

-------------------

* :class:`~scrapy_poet.page_input_providers.HttpRequestProvider`, added in
0.17.0, is now actually enabled by default.

-------------------

* Fixed ``scrapy savefixture`` not finding page object modules when used
outside a Scrapy project.

-------------------

* Now requires ``web-poet >= 0.17.0`` and ``time_machine >= 2.7.1``.

* Removed ``scrapy_poet.AnnotatedResult``, use
:class:`web_poet.annotated.AnnotatedInstance` instead.

* Added support for annotated dependencies to the ``scrapy savefixture``
command.

* Test improvements.

-------------------

* Added a ``.weak_cache`` to :class:`scrapy_poet.injection.Injector` which
stores instances created by providers as long as the :class:`scrapy.Request
<scrapy.http.Request>` exists.

* Fixed the incorrect value of ``downloader/response_count`` in the stats due
to additional counting of :class:`scrapy_poet.api.DummyResponse`.

* Fixed the detection of :class:`scrapy_poet.api.DummyResponse` when some type
hints are annotated using strings.