Re: [pkix] [apps-discuss] character repertoire for fragment identifiers

[Sam Ruby <[e-mail]>]

On 01/11/2015 05:44 PM, Sean Leonard wrote:
> On 1/11/2015 2:24 PM, Sam Ruby wrote:
>>
>> The next question is: are all possible URIs valid GeneralNames?
>>
>> Here are two examples:
>>
>> data:image/gif;base64,R0lGODlhyAAiALM...DfD0QAADs=
>
> Yes. Use case: Certificate Image. RFC 6170 (previously RFC 3709).
>
>> https://www.ietf.org/#related
>
> Yes. Use case: Certification Practice Statement (CPS) Pointer
> (Certificate Policies, id-qt-cps). Section 4.2.1.4 of RFC 5280. I have
> seen fragments in there in the wild.
>
> Both examples are valid.
>
> Note: the examples I provide here are actually examples where the URIs
> are encoded directly in IA5Strings; there is no GeneralName CHOICE. But
> the premise still stands. Others with PKIX experience can point to
> systems/protocols where uniformResourceIdentifier is in use with the
> appropriate.
>
> One offhand example that I was able to find, via a Google search, is the
> GS1 Certificate Profile Standard v2.0
> <http://www.gs1.org/gsmp/kc/epcglobal/cert/cert_2_0-standard-20100610.pdf>.
> That happens to use GeneralName. It's a bit elliptical, and refers to
> RFC 4043 (which I rarely see), but...it's there.

Thanks!  This input is helpful.  I don't see where it clearly states 
that fragments are allowed, but then again, I don't see where it says 
that they would be disallowed.  That coupled with your statement that 
you have seen them in practice (however rare) is valuable input.

> Sean

- Sam Ruby