Fixes#1997, #2068.
This is tricky because we need to get the return values right (see #1960 for
prior attempts) and we often run into python/mypy#3644. I found that I
could express most signatures correctly using a series of overloads.
A few other changes in here:
- Added splitunc, which according to https://docs.python.org/3/library/os.path.html
should exist in both Unix and Windows.
- Made the second argument to os.path.curdir Optional to match the implementation.
- Fixed os.path.split, whose previous Path-aware signature triggered python/mypy#3644.
Fixes#1961.
I mostly just replaced all str annotations with Text, including in return types. This is
only broadly correct; diffing a str and a unicode sequence actually results in a mixed
output of str and unicode. We could also keep the return types as str if using Text
causes errors in real code. For callbacks that take str, I introduced a Union alias
because a callable taking a str would not be a compatible with a parameter of type
Callable[[Text], bool].
I also fixed the return type of difflib.restore.
This fixes an error in Travis that seems to have been caused by python/mypy#4319.
The fix was taken from the stdlib/3.4/enum.pyi stub. Mypy no longer assumes
that classes whose metaclass is EnumMeta are subclasses of Enum, so we can't
bound the typevar on Enum.
As promised in #2014.
There are virtually no real changes between Python 2 and 3 in this module, but the stub had accumulated some meaningless differences. I also fixed a few incorrect types.
This is in our style guide. This is the last piece of offending code; I just submitted ambv/flake8-pyi#10 to enforce the rule in the linter in the future.
This was introduced in a recent version (so it doesn't affect us right now but will if we upgrade the flake8 version used in tests). It gives errors like:
```
/home/travis/build/ambv/flake8-pyi/.tox/typeshed/tmp/typeshed/stdlib/3/re.pyi:21:9: E741 ambiguous variable name 'I'
/home/travis/build/ambv/flake8-pyi/.tox/typeshed/tmp/typeshed/stdlib/3/re.pyi:39:5: E741 ambiguous variable name 'I'
/home/travis/build/ambv/flake8-pyi/.tox/typeshed/tmp/typeshed/stdlib/3/re.pyi:58:5: E741 ambiguous variable name 'I'
/home/travis/build/ambv/flake8-pyi/.tox/typeshed/tmp/typeshed/third_party/2/OpenSSL/crypto.pyi:43:5: E741 ambiguous variable name 'O'
```
These aren't useful for typeshed, because regardless of whether it's good style, these libraries really do contain variables named `I` and `O`.
- Made deque.maxlen read-only
- We don't support 3.2, so we don't care about signature changes in it
- There don't seem to be any missing set operations (I compared the dir() of this class to that of builtins.set)
This removes the circular dependency between the os and posix stub, which
is somehow triggering python/mypy#4442. We should ideally fix the mypy bug,
but since it's easy enough to fix the import cycle, we might as well do that
too.
I disagree with the recommendation that users create incomplete stubs, because such stubs lead to false positive type checker errors (see for example #525). I would like to instead set a norm that all stubs should contain all public objects in the library they cover.
RawConfigParser.read has code explicitly supporting PathLike objects.
Also removed an unused import and widened the accepted type from
Sequence to Iterable.
In Python 3 (but not Python 2), `object().__dir__()` works and returns a list of strings.
This is relevant when implementing a custom `__dir__` that invokes `super().__dir__()`.
These stubs are identical in Python 2 and 3, but I believe they
should not be merged, because there are numerous other modules
in the encodings package, and some only exist in some Python
versions. I don't think we can support that in the 2and3
directory.
* os: merge the top and bottom of os/__init__.pyi
Part of #1427. In preparation for merging the two stubs, I'm making the files
identical as much as possible. This PR merges the top of the file, down to
but not including the definition of statvfs_result, and the bottom up to
and including os.utime.
This PR mostly adds more "if sys.version_info" block. Until the merger
completes, we'll have some Python 2 blocks in the Python 3 stub and vice versa.
I also add a few missing constants and arguments.
In followup PRs I'll merge the rest of the file. I'll put the trickiest part
(the return values of functions like os.stat) in its own PR.
* back out DirEntry from py2
It relies on stat_result which we don't have yet in py2.
* urlparse: allow unicode arguments in more places
Part of #1411
urlunsplit() and similar functions accept either unicode or str in all
places. Their actual return type is unicode if any of the arguments is
unicode and nonempty, which the type system can't exactly express. I
left the return type as str because str is implicitly promoted to
unicode, so using the return type in a place that accepts unicode should
work.
unquote, parse_qs, and parse_qsl return unicode if you pass them unicode,
so AnyStr is appropriate in their stubs.
* fix type for urldefrag
The return type was wrong; this function returns a 2-tuple. The second member of the tuple is always a `str` if the argument type does not contain '#', and otherwise matches the type of the argument.
Jelle Zijlstra