Re: [art] [Last-Call] Artart last call review of draft-ietf-6man-rfc6874bis-02

[Brian E Carpenter <[e-mail]>]

On 19-Mar-23 10:51, Rob Sayre wrote:
> On Sat, Mar 18, 2023 at 2:18 PM Rob Sayre <[e-mail] <mailto:[e-mail]>> wrote:
> 
>     On Sat, Mar 18, 2023 at 1:55 PM Brian E Carpenter <[e-mail] <mailto:[e-mail]>> wrote:
> 
> 
>         The point of doing it the simple way is to allow users to use a simple law-of-least-astonishment cut and paste of a complete IPv6 address into a straightforward URL. The "IPvFuture" syntax was considered and rejected by the WG, as noted in the draft.
> 
> 
>     This part of your message doesn't make sense to me. Can't you use a tilde or something?
> 
> 
> Also, the WG is wrong. I know this is kind of rude to say, but that possibility must be considered, otherwise we wouldn't have last-calls. URLs are not straightforwardly cut-and-pasted, so this ideal is not really something that actually happens to users.

Except when it does. https://bugzilla.mozilla.org/show_bug.cgi?id=1759739 is
precisely an example of this. Regardless, the use cases we give are real.

> 
> You can find one instance of the issue here:
> https://github.com/whatwg/url/issues/150 <https://github.com/whatwg/url/issues/150>

That is isn't about the host part of a URI. By experiment, browsers are inconsistent
about percent-decoding of IPv6 literals in the host part.
  
> That is even about very unintuitive percent-encoding. But the WG is breaking the ugly machine-readable format. Don't do that, it is an Internet Standard.
> 
> Anyway, my feedback is that you should use the "IPvFuture" production, and this change is bad. I guess we don't need to go back and forth further.

Agreed.

Regards
     Brian
> 
> thanks,
> Rob
>