there is no year 0 in the Gregorian calendar

[karenetheridge]

This test in date.json is invalid and should be removed:

            {
                "description": "year 0000 is a leap year (0 % 400 == 0)",
                "comment": "https://www.rfc-editor.org/rfc/rfc3339#appendix-C — year zero leap year edge case",
                "data": "0000-02-29",
                "valid": true
            },

There is no year zero in the Gregorian calendar.

It looks like this was added via PR #884 (@vtushar06).

[vtushar06]

Thanks for catching this. You're right that the Gregorian calendar has no year zero.

For context on how it got in: RFC 3339's grammar allows 0000 (date-fullyear = 4DIGIT), §1 states the range as "0000AD and 9999AD", and the Appendix C leap_year formula returns true for year 0 - so the RFC's own machinery treats 0000-02-29 as a leap date. But that's the proleptic/astronomical reading, and it conflicts with the actual Gregorian calendar you're pointing at.

Since the two readings genuinely disagree (and python-jsonschema rejects year 0000 via datetime. MINYEAR while the grammar permits it), this looks like the same situation as the leading-zero discussion in #840 - an ambiguous case that's better excluded than asserted.