start should be optional: https://docs.python.org/3.8/library/itertools.html#itertools.count
Directly changing the second overload to mark start as optional
generated a mypy error about overlapping overloads, but adding a third
overload seems to work.
I also noticed that `_NStep` was defined as a TypeVar when I think it
makes more sense for it to be a Union, so I went ahead and changed that
as well.
The documentation says, "The parse() function can take either a filename
or an open file object."
(https://docs.python.org/3.8/library/xml.dom.minidom.html). The function
was annotated as only accepting a str, so I added IO[Any].
For now, the test requires changes that have been merged into the pytype master branch but not yet released. I'll update requirements-tests-py3.txt again once I've cut a new release (hopefully later this week).
* Fix pytype_test.
* Re-enable pytype_test.
* Remove continue-on-error from pytype_test.
* Temporarily install pytype from its master branch, as the code to
support the new directory structure hasn't been released yet.
* Set TYPESHED_HOME before using pytype.pytd.Typeshed().
I noticed that the pytype parse test was getting quite slow again. This
is a somewhat hacky change that caches pytype's internal pyi loader to
avoid parsing files multiple times. On my machine, the test goes from
taking ~6m to complete to ~30s.
See https://github.com/google/pytype/issues/779.
In Python 3.8, pytype fails to parse this file because it can't find
Context. This wasn't caught by pytype_test because the test runs in
Python 3.6.
Change some tkinter function parameter types from bool to int.
Based on the fact that tk.{YES,NO,TRUE,FALSE} are defined as ints
and that the usage example in the source code also uses an int as a bool.
* Add a version guard for importing enum.
I'm trying to pull the latest version of typeshed into Google, and
pytype chokes on pstats in Python 2 because the enum module was
introduced in 3.4.
* Move the enum import into the branch that defines SortKey.
The latest pytype release fixes two pyi parser bugs that allow files
affected by them to be taken off the pytype exclude list. I removed two
`total=False` declarations in tkinter/__init__ that pytype does not like
(because it checks that `total` is present only when TypedDict is a
class's immediate parent) and which shouldn't be needed because
_InMiscNonTotal already specifies totality. I double-checked that mypy
reports no errors in 3.7 on a .py file containing:
from typing_extensions import TypedDict
Foo = TypedDict('Foo', {'x': int}, total=False)
class Bar(Foo): pass
x: Foo = {}
showing that it doesn't require `total` to be repeated.
This attribute was removed in
https://github.com/python/typeshed/pull/3879 because it is undocumented
and was annotated incorrectly. Unfortunately, a surprising (?) amount of
Google code uses this attribute and assumes it is a deque, so it needs
to exist but can't have a more general annotation like Collection[Any].
I also removed a no-longer-used typing.Deque import.
This is more faithful to the implementation:
https://github.com/python/cpython/blob/3.5/Lib/datetime.py.
When these classes define __init__, pytype has trouble type-checking
classes that inherit from datetime.datetime (done in, e.g., the third party
datetime_tz library) because it gets confused about what arguments the
constructor expects.
This is a small cosmetic change. I want to encourage use of the
nicer-lookng pytype.config.Options.create() rather than
direct construction of the command-line args, so I'm changing
over all occurrences of the latter that I can find.
The main purpose of this change is to get the stubs into a
form that pytype will support with its next release. Since
these stubs already mostly use fully qualified names, this
will also increase consistency.
This also makes the flask stubs that depend on click parseable.
* In pytype_test, only parse files with the .pyi extension.
(There's a README in third_party/2and3/click/.)
* Annotate with ContextManager rather than using the
contextmanager decorator. This is shorter, removes a
dependency, and gets rid of the slight weirdness of a
decorator needing to be evaluated in a stub.
* Fix non-stub things.
During the PyCon sprints, I'm planning to bash the pytype pyi
parser into shape so we can handle all of third_party. As a first
step, create an explicit blacklist of everything pytype chokes on.
With today's release, pytype is able to fully load
stdlib/3/collections/__init__.pyi, so the test no longer needs the
ability to partially parse stubs using the pytd tool. Removing this
functionality allows the test code to be simplified considerably.
Speeds up pytype considerably by directly calling pytype.io.parse_pyi
rather than running `pytype --parse-pyi` in a subprocess.
The subprocess logic is still required to handle the one pyi file
marked `# parse only` in the pytype blacklist. Once
https://github.com/google/pytype/issues/242 is fixed, the test code
can be cleaned up considerably.
* Moves the pytype installation to requirements-tests-py3.txt, now
that pytype can run under Python 3.5+.
* Changes tests/pytype_test.py to not require a --python{version}-exe
argument when it can automatically find the Python interpreter, and
cleans up a few typos and out-of-date things.
* Updates the appropriate documentation.
* Updates .travis.yml.
* Add parts of third_party/ to pytype_test.
Pytype uses six and mypy_extensions, so these pyi files should be
tested with pytype.
* Remove newly loadable pyi files from pytype_blacklist.
