- From: Roy T. Fielding <fielding@gbiv.com>
- Date: Thu, 14 Jul 2011 19:07:13 -0700
- To: Marc Portier <mpo@outerthought.org>
- Cc: uri@w3.org
This is an old comment that I missed ...
On Apr 28, 2011, at 12:17 AM, Marc Portier wrote:
> Hi all,
>
> Finally got round at this javascript implementation of
> http://tools.ietf.org/html/draft-gregorio-uritemplate-04
>
> (minus the partial syntax and some more testing)
>
>
> Anyway, it is open at:
> https://github.com/marc-portier/uri-templates
>
> It wraps inside JQuery and comes with tests up in QUnit.
> (JQuery dependency currently limited to the handy $.extend and $.isFunction)
>
>
> Must say getting my head around things asked more time then I initially expected: clarity and elegance of the syntax and explanation did hide quite some more special cases then I would of have thought of.
>
> One case I still haven't ironed out, to the extend I'm liking my solution more :) is this:
>
>> tested pattern: {;list}
>> Expected: ";val1,val2,val3"
>> Result: ";list=val1,val2,val3"
>>
>> tested pattern: {;keys}
>> Expected: ";key1,val1,key2,val2"
>> Result: ";keys=key1,val1,key2,val2"
>>
>
> which in my mind (and implementation) follows more closely the line of thinking in:
>
>>
>> tested pattern: {?list}
>> Expected: "?list=val1,val2,val3"
>>
>> tested pattern: {?keys}
>> Expected: "?keys=key1,val1,key2,val2"
>>
>
> the consistency I see is this: both ; and ? operators deal with named parameters, in unexploded form the values are wrapped up in one string, but still expect value.
Yes, I made the same mistake in draft 05 for the Level 4 examples
but got it right in the section on Default. I'll fix it for 06.
....Roy
Received on Friday, 15 July 2011 02:07:39 UTC