dateutil first version #133
Conversation
| PERTAIN = ... # type: List[str] | ||
| TZOFFSET = ... # type: Dict[str, int] | ||
|
|
||
| def __init__(self, dayfirst: bool=..., yearfirst: bool=...) -> None: |
There was a problem hiding this comment.
Style suggestion: in stubs we prefer compactness and it's fine to put the "..." of the method body on the same line and omit the blank lines between methods.
|
Hey, sorry for the slow response. I just realized I've never used dateutil so I don't really have any idea whether this stub is adequate or not. Also, because it's third party, could you ask the dateutil author whether they are okay with having a dateutil stub in typeshed? See the note in README.md under "third_party". |
|
The original author of dateutil does not currently maintain the library, but I'm one of the two principal maintainers (and since the other guy has a newborn, probably the one most likely to respond to you) and I'm fine with a stub being included. |
|
|
||
| def parse( | ||
| self, | ||
| timestr: str, |
There was a problem hiding this comment.
I'm not overly familiar with how type annotations work, but you may want to be aware that timestr duck-types the input. In Python 3.x will take bytes, str or stream objects that implement read(). The preferred input is probably str, though.
There was a problem hiding this comment.
Maybe a union of bytes, str and IO would work.
|
@jukebox: I don't understand what happened, but it looks like a whole bunch of merged commits ended up in your PR. Could you clean that up (maybe try a rebase) or start a new PR? The comment from #133 (comment) indicates we can go ahead with this PR, but it's currently impossible to review due to the merge noise. |
|
is it ok now ? |
|
Yup, thanks! Does dateutil support Python 2 also? |
|
|
|
@jukebox Are you going to continue filling out the rest of the stub? If not, I am probably in a pretty good position to flesh out most of the library. I'll just need a bit of time to familiarize myself with the exact notation and whatnot. It seems that these stubs also cover the private interface, which makes sense, but makes things a bit more complicated, I would think, especially given #153. |
|
@pganssle i think i dont have time to continue the stub and i dont use other methods and old python versions. |
|
OK, I'll make a separate PR covering 2.x and expanding coverage to the whole library. |
No description provided.