OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

xri message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [xri] Backtracking

On 26-Oct-07, at 11:35 AM, Victor Grey wrote:

> Tan, William wrote:
>> 1. Authority resolution and service selection are distinct phases.  
>> When authority resolution completes, it is done and returns the  
>> result to the calling function. It might be at the 3rd Ref out of  
>> 5 possible Refs, and there is no way to continue where it left off.
>> 2. During service selection, if you see a Ref, you will spawn a  
>> different authority resolution phase, but that has nothing to do  
>> with the original authority resolution. This child authority  
>> resolution phase also knows nothing about service selection. It  
>> completes, and returns (if successful) the results back to the  
>> caller, which is the service selection function. If this child  
>> authority resolution phase fails and there are more Refs, the  
>> service selection function will spawn another authority resolution  
>> phase on the next Ref.
> Yes, well said, and that matches my understanding.
> John Bradley wrote:
>> This is the tricky bit.  I am informed that the current spec sec  
>> 11.5 intends that if the final XRD is returned and you do service  
>> selection on it,  and if that "does not" find the required SEP  
>> then you must unwind back to the same point that you would have if  
>> the XRD was not returned.
> So, are you saying that if I resolve @a*b*c and the XRD for *c does  
> not contain the SEP I'm looking for, nor any REFs, I should go back  
> and look in *b for that SEP? That would not make any sense to me.

My point exactly but in the spec says go back to *b and continue with  
the next ref in priority order.

> If however I follow an XRD level REF that I found in *c, and do not  
> find the SEP I'm looking for in the last XRD for -that- resolution  
> chain, I can go and examine any other REFs that I might have found  
> in *c. But once I start following REFs, I cannot backtrack to  
> before the XRD where I started following REFs.
Another issue I have raised is that if a XRD has a ref at the XRD  
level no SEPS are processed in that XRD.  So if by other refs in *c  
you mean a ref in the default service then yes.

> =vg


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]