ABNF for sf-date

[Jxck]

according to "Parsing a Date"

https://github.com/httpwg/http-extensions/blob/main/draft-ietf-httpbis-sfbis.md#parsing-a-date-parse-date

Let output_date be the result of running Parsing an Integer or Decimal ({{parse-number}}) with input_string.

align ABNF to this line, I think sf-date should be like below

- sf-date = "@" ["-"] 1*15DIGIT
+ sf-date = "@" sf-integer

[reschke]

Not convinced. It adds a level of indirection; that is, it makes the ABNF less readable.

[martinthomson]

Indirection means DRY, which edges out that sort of readability for me.

[LPardue]

I find the ANF confusing anyway. The prose says parse it as Integer or Decimal, then If output_date is a Decimal, fail parsing. For a library that cares to differentiate between a parse failure due to bad characters vs bad types, then the ABNF would imply that you're not allowed to parse the Decimal type correctly?

[Jxck]

@LPardue IIUC, that's why this draft made ABNF as Appendix.
That caused since Algorithm has "Parse Number" returns Integer | Decimal, but not in ABNF.
What should be intends is just allow sf-integer, and in serialization point of view, it works fine.

[mnot]

I don't have particularly strong feelings about this; does anyone else?

[LPardue]

The ABNF already has layers upon layers of indirection. I think trying to avoid it for this field is inconsistent. So I'm sort of with MT that DRY wins out.