Test the docs on Python 2 and 3 using manuel. This gets a lot of coverage of tests.attrhooks, tests.cucumbers and tests.utils.

Build CFFI modules the modern way.

Switch to zope.testrunner from 'python setup.py test' and 'nose'.

Both are essentially deprecated.

This is a step in producing usable coverage numbers.

(Also remove cp33 from the manylinux build. This was missed in https://github.com/zopefoundation/persistent/pull/74)

Add support for Python 3.7 and drop 3.3

[skip ci]

[skip ci]

Update url in setup.py

Fixes #71

Also fix the lint warnings in setup.py: consistent use of continued indentation and trailing commas.

Change cPickleCache to use PER_TypeCheck

Instead of testing object sizes.

This matches what the pure Python implementation does and is a
stronger test that the object really is compatible with the cache.
Previously, an object could potentially include ``cPersistent_HEAD``
and *not* set ``tp_base`` to ``cPersistenceCAPI->pertype`` and still
be eligible for the pickle cache; that is no longer the case.

This resolves several compiler warnings:

persistent/cPickleCache.c:521:43: warning: comparison of integers of different signs: 'Py_ssize_t' (aka 'long') and 'unsigned long' [-Wsign-compare]
                (v->ob_type->tp_basicsize >= sizeof(cPersistentObject))
                 ~~~~~~~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~~~~~~~~~~~
persistent/cPickleCache.c:709:39: warning: comparison of integers of different signs: 'Py_ssize_t' (aka 'long') and 'unsigned long' [-Wsign-compare]
    else if (v->ob_type->tp_basicsize < sizeof(cPersistentObject))

Fixes #69.

Update documentation links.

Added a link to readable documentation, viewable from GitHub.

Fix possibility of rare crash during dealloc. Fixes #66

remove build files in manylinux builds

CI setup for manylinux wheels on tagged commit

Avoid raising a SystemError when clearing slots if setstate() failed.

PR #52 introduced a code path to `ghostify` that calls PyErr_Clear()
with the intent to avoid propagating AttributeErrors for slots.

However, if there is an error (like a POSKeyError) raised by
jar.setstate(), then `unghostify` will call ghostify with an error
pending. If the object had slots that weren't set and the
AttributeError was cleared, so was the pending error from setstate. So
when `ghostify` returned NULL that got propagated up to the
interpreter which finds no exception and so raises `SystemError: error
return without exception set`.

This commit makes `unghostify` save and restore the exception state
around the call to PyErr_Clear.