adding methods for yearMonth and MonthDay

[jasonwilliams]

Adds methods to yearMonth and monthDay

[nekevss]

Looks good to me whenever you want to take it off draft.

The one thing missing is maybe the default implementation and some unit tests, but I don't think those are necessarily needed to merge 😄

[jasonwilliams]

@nekevss let me know your thoughts

I've been trying to add add() but it seems it's iso is using the date version of add_date_duration, but it seems like i need an implementation more like dateTime's:

temporal/src/iso.rs

Lines 125 to 164 in dd51ebb

	/// Specification equivalent to 5.5.9 `AddDateTime`.
	pub(crate) fn add_date_duration<C: CalendarProtocol>(
	&self,
	calendar: &CalendarSlot<C>,
	date_duration: &DateDuration,
	norm: NormalizedTimeDuration,
	overflow: Option<ArithmeticOverflow>,
	context: &mut C::Context,
	) -> TemporalResult<Self> {
	// 1. Assert: IsValidISODate(year, month, day) is true.
	// 2. Assert: ISODateTimeWithinLimits(year, month, day, hour, minute, second, millisecond, microsecond, nanosecond) is true.
	// 3. Let timeResult be AddTime(hour, minute, second, millisecond, microsecond, nanosecond, norm).
	let t_result = self.time.add(norm);
	// 4. Let datePart be ! CreateTemporalDate(year, month, day, calendarRec.[[Receiver]]).
	let date = Date::new_unchecked(self.date, calendar.clone());
	// 5. Let dateDuration be ? CreateTemporalDuration(years, months, weeks, days + timeResult.[[Days]], 0, 0, 0, 0, 0, 0).
	let duration = Duration::new(
	date_duration.years,
	date_duration.months,
	date_duration.weeks,
	date_duration.days + f64::from(t_result.0),
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	)?;
	// 6. Let addedDate be ? AddDate(calendarRec, datePart, dateDuration, options).
	let added_date = date.add_date(&duration, overflow, context)?;
	// 7. Return ISO Date-Time Record { [[Year]]: addedDate.[[ISOYear]], [[Month]]: addedDate.[[ISOMonth]],
	// [[Day]]: addedDate.[[ISODay]], [[Hour]]: timeResult.[[Hour]], [[Minute]]: timeResult.[[Minute]],
	// [[Second]]: timeResult.[[Second]], [[Millisecond]]: timeResult.[[Millisecond]],
	// [[Microsecond]]: timeResult.[[Microsecond]], [[Nanosecond]]: timeResult.[[Nanosecond]] }.
	Ok(Self::new_unchecked(added_date.iso, t_result.1))
	}

The former doesn't use the calendar at all by the looks of it

secondly why does the leap_year not return a bool? Is that wasteful?

temporal/src/utils.rs

Lines 230 to 233 in dd51ebb

	/// Returns either 1 (true) or 0 (false)
	pub(crate) fn mathematical_in_leap_year(t: f64) -> i32 {
	mathematical_days_in_year(epoch_time_to_epoch_year(t)) - 365
	}

[nekevss]

Here's some feedback regarding the most recent changes 😄

I don't think YearMonth or MonthDay call add or subtract directly. It's mostly done with Date in the middle

[nekevss]

There's one last thing that I missed the last couple times around. But outside of that this is overall looking fantastic!

[nekevss]

Looks great!

[jedel1043]

Oh, I had an unsent pending review 😅 I'll push the comment and I think we can address it on a separate PR

[jedel1043]

Why do we need to expose the fields as public?
Also, I think this should be a proper struct with named fields for clarity