Format Time#inspect with Internet Extended Date/Time Format

[straight-shoota]

The Internet Extended Date/Time Format (IXDTF) as described in RFC 9557 is a standard for representing a timestamp together with zone information.

Changes to the format:

• Replace UTC location by Z offset
• Skip zero nanoseconds entirely even when with_nanoseconds is true (see Format Time#inspect with Internet Extended Date/Time Format #16039 (comment))
• Remove whitespace between time and offset
• Wrap location in square brackets to indicate an IXDTF time-zone suffix

Resolves #15810

[ysbaddaden]

Do we have support for parsing the format?

[straight-shoota]

Not yet. But that's the follow-up.

[straight-shoota]

thought: Dropping the second fractions entirely when nanoseconds are zero might be debatable. I think it's more readable that way.
But that makes it ambiguous whether nanoseconds were entirely cut off (with_nanoseconds: false) or are just omitted because the subsection fraction is zero. Shortening the fraction to just .0 was more explicit. But I don't think it's that important.
with_nanoseconds: false is a rather odd feature. I'm not aware of any actual use case. It was introduced in #5794 but I don't recall what was the reason for it.

[straight-shoota]

We should probably deprecate with_nanoseconds.

[straight-shoota]

Sorry for the continued editing but I realized that essential parts were missing (documentation) and the exact format could use some polishing for better readability.

I have now replaced the implementation with the existing Format::RFC3339 formatter, but keeping space as separator instead of T which is harder to make out. RFC3339 explicitly allows that to improve human readability, which is the main purpose of this method.

[ysbaddaden]

Moving to a RFC or ISO standard format made sense, but not using T for readability reasons makes me wonder about the change because the most readable is the current format (SP TZ > [TZ] and SP UTC > Z).

I guess the question is: is #inspect meant for humans, in which case it must be the best readable, or for machines, in which case a standard format is better.

[straight-shoota]

Yes, Object#inspect is primarily meant for human consumption. It doesn't hurt to have it machine readable as much as possible, of course. But that's not a strict goal.

Anyways, using a space instead of T between date and time is covered by the specification as far as I understand. It's not the canonical format, but applications are allowed to choose it for readability purposes. I believe the rationale here is that RFC 3339 and IXDTF primarily focus on defining a formal standard format for machine consumption and don't worry too much about readability.
But some allowances are possible when the format is intended for a human audience.

Stale