Check for leap year when calculating differenceInMonths involving February

[kaunartist]

Problem

When calling differenceInMonths from January 31st to February 28th on a leap year, the logic for handling February does not account for the 29th. As a result intervalToDuration returns a duration of 1 month, -1 day.

Example

> differenceInMonths(new Date(2023, 1, 28), new Date(2023, 0, 31))
1
> differenceInMonths(new Date(2024, 1, 28), new Date(2024, 0, 31))
1
> intervalToDuration({ start: new Date(2023, 0, 31), end: new Date(2023, 1, 28) })
{ months: 1 }
> intervalToDuration({ start: new Date(2024, 0, 31), end: new Date(2024, 1, 28) })
{ months: 1, days: -1 }
> intervalToDuration({ start: new Date(2024, 0, 31), end: new Date(2024, 1, 29) })
{ months: 1 }

Expected Output

> differenceInMonths(new Date(2023, 1, 28), new Date(2023, 0, 31))
1
> differenceInMonths(new Date(2024, 1, 28), new Date(2024, 0, 31))
0
> intervalToDuration({ start: new Date(2023, 0, 31), end: new Date(2023, 1, 28) })
{ months: 1 }
> intervalToDuration({ start: new Date(2024, 0, 31), end: new Date(2024, 1, 28) })
{ days: 28 }
> intervalToDuration({ start: new Date(2024, 0, 31), end: new Date(2024, 1, 29) })
{ months: 1 }

• Ran tests
• Ran linter

[BenediktKa]

This issue is also affecting formatDistance, when using 1st of February 2024 and 1st of March 2024. This is due to the dependence of the differenceInMonths() as we noticed today in one of our test suits.