- adds a base64url format description

[baywet]

related #3167

[handrews]

@baywet I think you need to update this to remove the deprecated version info.

I'm also wondering if, now that we have definitively settled on byte rather than base64, if it makes sense to align with RFC 4648 terminology for this, or call it byte-url or something. Neither feels very satisfactory to me.

[baywet]

Can you expand on why do you think we should remove the deprecated aspect please? I might have missed part of the conversation here.

If it were only up to me, byte would be base64 and base64url would be base64url. byte only being an alias to byte64 for compatibility reasons. Those different names are simply adding confusion IMHO.

[handrews]

Edit to my previous comment: formats are kebab-case so byte-url, not byteUrl

Re: deprecated_note — see #3167 (comment)

In #3174 I suggested having a base64 alias for byte but we discussed it in the last call and decided not to do that (as seen in #3182). So we're stuck with byte.

[baywet]

@handrews so are we good with this PR as is or do you suggest we make other changes?

[handrews]

I'm still feeling meh about whether to use base64url for compatibility with OAS 3.1 vs byte-url for compatibility with OAS 3.0, the actual targeted version here. As much as I dislike byte, it's what we're stuck with and this is a URL-safe version of it. Although I suspect that very few tools will implement support for this so I'm not sure how much it matters.

@MikeRalphson, @darrelmiller, anyone else - any preference for base64url vs byte-url in the context of OpenAPI 3.0?

[MikeRalphson]

+1 for base64url, let's not be held back by our past missteps.

[handrews]

@MikeRalphson thanks - in that case this looks good to go.

[MikeRalphson]

LGTM