Iterators for Ranges of Dates with day steps

[Nikita-str]

The issue: #5448

The PR implements iterators with day steps for Date ranges by next pipeline:
RataDie -> Iso::iso_from_fixed -> Date::new_from_iso

The PR is only partial solution for the issue because:

• There no month/year-step iterator: what to do when there no such month in next year or such day in next month? Round up to a bound?
• You can make iterations faster on small intervals of days for ArithmeticDate (at least if day step is less or equal than min possible days in a month of calendar)

[sffc]

Thanks for your interest!

The linked issue is about date range formatting, not necessarily date increment calculations. Now, adding durations to dates is something we want, and which we have partially implemented.

I would support a PR adding Date::next_date and Date::prev_date fns. Not sure about the iterator because the use cases seem fairly limited for the amount of extra code it adds.

[sffc]

Not sure about the iteration API; @Manishearth wdyt?

[Manishearth]

I don't want to add an entirely new API for this. A Step trait impl would be nice so that people can do for date in <start>..<end>, and that trait can be very easily implemented using the offset() function.

I do not see a large benefit of a dedicated API that lets you control the step size, especially since it only supports step sizes measured in days and no other unit.

[Nikita-str]

The linked issue is about date range formatting

lol

I would support a PR adding Date::next_date and Date::prev_date fns.

understood

A Step trait impl would be nice so that people can do for date in <start>..<end>

Hmm... but Step trait is unstable, or do you mean to reduce functionality of the Iters to single day step?

[Manishearth]

Oh, if it's unstable we should just not do this until it stabilizes.

[Nikita-str]

So, close the PR then?

[sffc]

If a dedicated pub fn next_date(&self) has characteristics that make it smaller/faster than pub fn add(&self, DateDuration), then we could consider adding next_date.

[sffc]

The linked issue is about date range formatting

lol

Sorry about the misleading issue!

In general, if you want to make contributions, consider issues labeled good first issue, which should generally have more detail on what the goals are.