Last Call: <draft-gregorio-uritemplate-07.txt> (URI Template)

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Last Call: <draft-gregorio-uritemplate-07.txt> (URI Template)

Frank Ellermann-3
Hi, that's an important and good draft.  Some editorial nits:

In section 2.1 you use CTL, DQUOTE, and SP in a comment.
Please add these terms to the ABNF imports in section 1.5.

In section 1.3 you mention WSDL, WADL and OpenSearch.
Please add informative references and expand the acronyms.

Please update the TUS reference to 6.x.  There are no
changes wrt the concepts used in this draft (stability of
non-characters, etc.), but I think UTR #15 is an integral
part of TUS since 5.0 (?)

In section 3.1 you write:

| If the literal character is allowed anywhere in the URI
| syntax (unreserved / reserved / pct-encoded ), then it is
| copied directly

Do you mean "is allowed in the given part of the URI" here?
What I have in mind are, e.g., %x5B and %x5D in a query or
fragment.  By definition in 2.1 these are "literals", but
have to be percent-encoded n STD 66 queries or fragments.

-Frank

Reply | Threaded
Open this post in threaded view
|

Re: Last Call: <draft-gregorio-uritemplate-07.txt> (URI Template)

Mark Nottingham-2
Hi Frank,

Thanks for the feedback. Responses below.

On 29/11/2011, at 8:23 PM, Frank Ellermann wrote:

> Hi, that's an important and good draft.  Some editorial nits:
>
> In section 2.1 you use CTL, DQUOTE, and SP in a comment.
> Please add these terms to the ABNF imports in section 1.5.

I'm -0 on this, as they're informative, whereas 1.5 is normative.


> In section 1.3 you mention WSDL, WADL and OpenSearch.
> Please add informative references and expand the acronyms.

In SVN.


> Please update the TUS reference to 6.x.  There are no
> changes wrt the concepts used in this draft (stability of
> non-characters, etc.), but I think UTR #15 is an integral
> part of TUS since 5.0 (?)

In SVN.


> In section 3.1 you write:
>
> | If the literal character is allowed anywhere in the URI
> | syntax (unreserved / reserved / pct-encoded ), then it is
> | copied directly
>
> Do you mean "is allowed in the given part of the URI" here?
> What I have in mind are, e.g., %x5B and %x5D in a query or
> fragment.  By definition in 2.1 these are "literals", but
> have to be percent-encoded n STD 66 queries or fragments.


Not sure what you're saying here; the URI escape syntax is % HEXDIG HEXDIG. If the literal string "%x5B" occurs in a template, it'll be copied into the result, since it looks like a percent-encoded ("%x5") followed by a "B". If the literal character "[" occurs in a template, it'll also be copied into the result, since that's part of reserved (thanks to gen-delims).

The intent here is definitely for a processor NOT to need to know what part of the URI it's in, since templates can make this ambiguous.

--
Mark Nottingham   http://www.mnot.net/




Reply | Threaded
Open this post in threaded view
|

Re: Last Call: <draft-gregorio-uritemplate-07.txt> (URI Template)

Frank Ellermann-3
On 30 November 2011 00:44, Mark Nottingham <[hidden email]> wrote:

> Not sure what you're saying here; the URI escape syntax is % HEXDIG HEXDIG.

ACK, sorry for the confusion, I used the same ABNF hex. constructs as in
the "literals" section for the square brackets.

> If the literal character "[" occurs in a template, it'll also be copied
> into the result, since that's part of reserved (thanks to gen-delims).

> The intent here is definitely for a processor NOT to need to know what
> part of the URI it's in, since templates can make this ambiguous.

Okay, that was answered my question, when the draft says "anywhere" it
actually means "anywhere", even if the output is no valid UEL.

-Frank