Skip to content
Merged
Changes from 3 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions stdlib/datetime.pyi
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@ class date:
@property
def day(self) -> int: ...
def ctime(self) -> str: ...
def strftime(self, fmt: str) -> str: ...
def strftime(self, format: str) -> str: ...

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
def strftime(self, format: str) -> str: ...
def strftime(self, __format: str) -> str: ...

Let's make it positional-only as discussed.

@karolyi karolyi Nov 18, 2021 •

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Quoting https://bugs.python.org/msg373421, specifically:

I think in typeshed they can safely change from fmt to format even today (which would almost certainly be more accurate to end user use cases).

What do you have against this?

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Type checking is all about warning users about situations where their code could lead to unexpected errors. Passing a keyword argument to this parameter will currently cause code to raise an error sometimes (but not all the time, or even most of the time), as not all systems can be guaranteed to use the C implementation of datetime (this is partly why the Python implementation exists alongside the C implementation). If a user passes a positional argument to this parameter, however, then there is no possibility of an error. So, it seems correct that, as the situation currently stands, a type-checker should raise an error if a keyword argument is passed to this parameter, as doing so could in some situations lead to unexpected errors.

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's definitely better than the current situation to have the parameter named format, but I agree that the best option for now is to have it as a positional-only parameter in the stub.

def __format__(self, fmt: str) -> str: ...
def isoformat(self) -> str: ...
def timetuple(self) -> struct_time: ...
Expand Down Expand Up @@ -118,7 +118,7 @@ class time:
if sys.version_info >= (3, 7):
@classmethod
def fromisoformat(cls: Type[_S], time_string: str) -> _S: ...
def strftime(self, fmt: str) -> str: ...
def strftime(self, format: str) -> str: ...

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
def strftime(self, format: str) -> str: ...
def strftime(self, __format: str) -> str: ...

def __format__(self, fmt: str) -> str: ...
def utcoffset(self) -> timedelta | None: ...
def tzname(self) -> str | None: ...
Expand Down