On Tue, Oct 19, 2021 at 8:54 PM Hongyi Zhao <[email protected]> wrote:
> On Wed, Oct 20, 2021 at 8:29 AM Robert Kern <[email protected]> wrote: > > > > On Tue, Oct 19, 2021 at 8:22 PM <[email protected]> wrote: > >> > >> Do I have to use it this way? > > > > > > Nothing is forcing you to, but everyone else will write it as > `dtype=bool`, not `dtype=(bool)`. `dtype=(bool)` is perfectly > syntactically-valid Python. It's just not idiomatic, so readers of your > code will wonder why you wrote it that way and if you meant something else > and will have trouble reading your code. > > I use Emacs, and find that python-lsp-server will give the dtype=() as > the completion result, see here [1] for more detailed discussion. > > [1] https://github.com/python-lsp/python-lsp-server/issues/98 Well, that's not helpful of python-lsp-server. Is it looking at our type annotations? The annotations for `dtype=` arguments are likely quite complicated because we accept so many different types here, but `()` is not a great completion. -- Robert Kern
_______________________________________________ NumPy-Discussion mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: [email protected]
