Re: Prefer Draft Feedback

On 2011-12-06 20:13, Alex Rousskov wrote:
> On 12/06/2011 11:43 AM, James Snell wrote:
>> On Tue, Dec 6, 2011 at 8:29 AM, Alex Rousskov wrote:
>>> Overall, the Prefer grammar is more complex than any similar construct I
>>> saw in HTTPbis Part 1 or Part 6! If I did not miss any existing cases,
>>> would it be possible to simplify the Prefer grammar so that existing
>>> parsing code and data structures can be reused to deal with it?
>
>> The grammar for Prefer is modeled closely after the Expect header
>> field grammar in Part 2 and adds only the allowance that a preference
>> can have it's own value (e.g. "Prefer: wait=10" ...
>
> As Julian pointed out, expectation-extension in RFC 2616 attempts to use
> the same syntax:
>
>>    Expect       = 1#expectation
>>
>>    expectation  = "100-continue" / expectation-extension
>>    expectation-extension = token [ "=" ( token / quoted-string )
>>                             *expect-params ]
>>    expect-params = ";" token [ "=" ( token / quoted-string ) ]

I just sanitized the grammar to:

   expectation  = "100-continue" / expectation-extension
   expectation-extension = token [ "=" ( token / quoted-string )
                            *(";" expect-param) ]
   expect-param = token [ "=" ( token / quoted-string ) ]

(<http://trac.tools.ietf.org/wg/httpbis/trac/changeset/1489>).

Also, I think while converting the ABNF, we forgot to allow OWS around 
";"; I opened a ticket for that: 
<http://trac.tools.ietf.org/wg/httpbis/trac/ticket/327>

> The only difference I see is that expect-params are only allowed for
> expectations that start with name=value but I bet that is just a bug
> with the closing ']' placement in expectation-extension definition
> (something for HTTBis to fix?).

Make sense to me; I don't believe anybody is using this anyway, so let's 
make Prefer and Expect at least the same.

(I added your observation to the ticket).

> If expectation-extension is fixed, we would have to support the same
> syntax for Expect, my "this is too complex!" comment would be
> essentially wrong, and you can have the nice-looking wait=10 preference.

Agreed.

I also need to get this pattern into the Wiki page 
<http://trac.tools.ietf.org/wg/httpbis/trac/wiki/HeaderFieldTypes> 
(contributions welcome!).

Best regards, Julian