Re: [apps-discuss] Potential issues in RFC 3986

[Sam Ruby <[e-mail]>]

On 01/02/2015 11:54 AM, Julian Reschke wrote:
> On 2015-01-02 16:58, Sam Ruby wrote:
>> ...
>>>> ...
>>>>> In a brief sampling, I couldn't see any divergence which is likely
>>>>> to be
>>>>> resolvable by changing the URI spec.
>>>>
>>>> I encourage you to spend more time with that data.  An example of a
>>>> concrete problem is handing of hosts in a UTS-46 compliant manner.
>>>> ...
>>>
>>> Which test, specifically?
>>
>> Most of the tests in the range of 242..263 deal with IDNA issues in some
>> manner or another.  A particularly fun one is test 261:
>>
>> https://url.spec.whatwg.org/interop/test-results/e99bce2c50
>>
>> <aside>ordinal test numbers may change over time.  The hash, however,
>> will not</aside>.
>>
>> I encourage you to scan the text of RFC 3986 for the term "IDNA".  In
>> the decade between when RFC 3986 was published and today, things have
>> progressed a bit here.
>> ...
>
> I'm fully aware that RFC 3986 is likely incorrect wrt IDNA advice.
>
> However, the test that you cite does not use a valid URI, so I don't
> think it can be used as example that something in RFC 3986 is incorrect.

Since you agree that RFC 3986 is likely incorrect wrt IDNA advice, can I 
get you to suggest a test case that demonstrates this?

The master source for these tests is here:

https://github.com/w3c/web-platform-tests/blob/master/url/urltestdata.txt

A pull request would be appreciated.  A detailed suggestion with the 
same information (a source string, a base URI, and what the expected 
values for each of the components) also works.

  - - -

Railing this up a level, Roy Fielding believes that a WG would need to 
be chartered in order to update RFC 3986.  See the bottom of the 
following email:

http://lists.w3.org/Archives/Public/public-ietf-w3c/2014Dec/0088.html

Per that email, I tend to agree.

On any specific issue, I don't know what the outcome should be -- 
whether RFC 3986 should be updated, or whether implementations should 
change.  What I am looking for is to get together the set of people who 
are interested in that discussion and for us to come to a consensus 
together.

I'm willing to do the leg-work, whether that be running tests or 
updating specs, or filing bugs against implementations.

> Best regards, Julian

- Sam Ruby