In its current form, the arg(generator=bool_with_no_arg) slightly conflicts with a few other plans I had in mind: Since the @overloads of the arg basically have an "exponential explosion" issue (basically every new feature switch requires to double the number of overloads), I'm considering to switch to more specific argument annotations, i.e., possibly different ones for plain T, nargs T, boolean switches, and perhaps also splitting it up for positional vs non-positional ones. After that change, having such special kind of boolean switches, and even fully custom / user-defined argument types would be much easier to support. I hope I can find some more time to really think this through (sorry for the delays in general -- just not finding enough time on this recently)!
In its current form, the
arg(generator=bool_with_no_arg)slightly conflicts with a few other plans I had in mind: Since the@overloads of theargbasically have an "exponential explosion" issue (basically every new feature switch requires to double the number of overloads), I'm considering to switch to more specific argument annotations, i.e., possibly different ones for plainT, nargsT, boolean switches, and perhaps also splitting it up for positional vs non-positional ones. After that change, having such special kind of boolean switches, and even fully custom / user-defined argument types would be much easier to support. I hope I can find some more time to really think this through (sorry for the delays in general -- just not finding enough time on this recently)!