Change formatting/marshaling of time.Duration

[aurelius-adrian]

Proposal:

Seeing that the duration of a month varies greatly depending on which month in which year, marshaling time.Duration into ISO8601 with a month component would be ambiguous. As far as I know there is no context contained within the time.Duration type, as it is saved as a duration in nano seconds (int64). Therefore time.Duration should be marshaled into ISO8601 without year and month components.

According to ISO8601 3.1.1.8
"...The exact duration of a time scale unit (3.1.1.7) depends on the time scale (3.1.1.5) used. For example, the durations of a year, month, week, day, hour or minute, may depend on when they occur [in a Gregorian calendar (3.1.1.19), a calendar month (3.1.2.19) can have a duration of 28, 29, 30, or 31 days; in a 24-hour clock (3.1.1.10), a clock minute (3.1.2.4) can have a duration of 59, 60, or 61 seconds, etc.]. Therefore, the exact duration can only be evaluated if the exact duration of each is known.
Note 4 to entry: This definition is closely related to NOTE 1 of the terminological entry “ duration ” in IEC 60050-113:2011, 113-01-13."

I understand that the ISO definition states that even the number of minutes within a hour can vary, I believe in practice most parsers assume a 60min hour, however parsers, at least those I require, do not assume a fixed duration in ns for the Month component - resulting in errors when unmarshaling time.Duration.

The imported library:

gqlgen/graphql/duration.go

Line 7 in 7b1908e

dur "github.com/sosodev/duration"

which is used here:

gqlgen/graphql/duration.go

Line 26 in 7b1908e

return MarshalString(dur.Format(d))

sets a fixed month component duration here:
https://github.com/sosodev/duration/blob/b7258bf670d35309ff417d0ab25fc77f9d338cfa/duration.go#L40