Re: [xri] Removing priority attribute from XRD altogether

[Joseph Anthony Pasquale Holsten <joseph.holsten@cordance.net>]

It's an XPath that matches every Link element which contains at least  
one Rel element with the contents "stylesheet". So a good XPath  
implementation should be able to do XRD related resource processing  
with something like:

   def xrdRelProcess(doc, rel)
     rel = escapeXML(rel)
     xpath = sprintf('/xrd:XRD/xrd:Link[Rel="%s"]', rel)
     namespaces = {'xrd' => "http://docs.oasis-open.org/ns/xri/xrd-1.0"}
     links = doc.xpath(xpath, namespaces)
     return links
   end

Getting all the URIs and URITemplates from Links with a matching Rel  
is a bit more awkward, but still a one liner:
   /xrd:XRD/xrd:Link[xrd:Rel="stylesheet"]/child::*[self::xrd:URI |  
self::xrd:URITemplate]
Again, that's just matching for stylesheet links. You get the idea.

If we don't use document ordering, not only can I not handle this in a  
1-liner, I can't seem to do it in just XPath alone. So you'd need to  
pull out XQuery:

   for $uri in /xrd:XRD/xrd:Link[xrd:Rel="stylesheet"]/ 
child::*[self::xrd:URI | self::xrd:URITemplate]
   order by $uri/../@Priority $uri/@Priority
   return $uri

I think. I don't have an XQuery implementation lying around, so that's  
mostly a guess.

But in investigating implementations, I've found a little more trouble  
than I'd hoped for. The main problem with this is that libxml2 doesn't  
properly handle XPath expressions which include XML entities. That  
means it can't handle the right XPath for something like  
<Rel>foo'bar"baz&quux>zip<zap</Rel>, which should be /xrd:XRD/ 
xrd:Link[Rel="foo&apos;bar&quot;baz&amp;quux&gt;zip&lt;zap"] or at  
least /xrd:XRD/xrd:Link[Rel="oo'bar&quot;baz&quux>zip<zap"]. Since  
libxml2 is behind a ton of implementations, I'm really trying to fix  
this bug this week.

Of course, if you grok all this in one read I envy your XPath-fu.



On Aug 31, 2009, at 12:08 AM, Drummond Reed wrote:

> Joesph, maybe I've just spent too much time remodelling this  
> weekend, but
> can you unpack for me why /xrd:XRD/xrd:Link[xrd:Rel="stylesheet"] is  
> pretty
> spiffy?
>
> Thanks,
>
> =Drummond
>
>> -----Original Message-----
>> From: Joseph Anthony Pasquale Holsten [mailto:joseph.holsten@cordance.net 
>> ]
>> Sent: Saturday, August 29, 2009 8:17 PM
>> To: XRI TC
>> Subject: Re: [xri] Removing priority attribute from XRD altogether
>>
>> +1 because /xrd:XRD/xrd:Link[xrd:Rel="stylesheet"] is pretty spiffy.
>>
>> On Aug 29, 2009, at 8:24 PM, John Bradley wrote:
>>
>>> We lose a round robin load balancing that no one ever used.
>>>
>>> Some things that don't preserve order may break.
>>>
>>> The addition of dsig requires order to be preserved anyway.
>>>
>>> I vote for dropping the priority attribute to simplify things.
>>>
>>> John B.
>>> On 29-Aug-09, at 9:15 PM, Gabe Wachob wrote:
>>>
>>>> Wow.
>>>>
>>>> This feels like deja vu, only the consensus on this list is 180
>>>> degrees different.
>>>>
>>>> I'm all for relying on ordering, always have been. Any There was
>>>> significant opposition
>>>>
>>>> +1e6 to dropping the priority attribute.
>>>>
>>>> -Gabe (yes, I'm still a TC member for the time being)
>>>>
>>>> On Fri, Aug 28, 2009 at 10:25 AM, <njones@ouno.com> wrote:
>>>>> +1 On dropping the priority attribute.
>>>>>
>>>>> On Thu, 27 Aug 2009 22:07:26 -0700, Eran Hammer-Lahav
>> <eran@hueniverse.com
>>>>>>
>>>>> wrote:
>>>>>> If the XML experts on the list feel comfortable with the  
>>>>>> ability to
>>>>>> preserve document ordering, I am very much in favor of dropping  
>>>>>> the
>>>>>> priority attribute.
>>>>>>
>>>>>> EHL
>>>>>>
>>>>>>> -----Original Message-----
>>>>>>> From: Scott Cantor [mailto:cantor.2@osu.edu]
>>>>>>> Sent: Thursday, August 27, 2009 6:30 PM
>>>>>>> To: 'Peter Davis'
>>>>>>> Cc: 'XRI TC'; 'DeWitt Clinton'
>>>>>>> Subject: RE: [xri] Removing priority attribute from XRD  
>>>>>>> altogether
>>>>>>>
>>>>>>> Peter Davis wrote on 2009-08-27:
>>>>>>>> My concern with this is that i fear that XRD instances which  
>>>>>>>> are
>>>>>>>> dynamically generated within an application are able to produce
>>>>>>>> service elements in the order they intend when serializing  
>>>>>>>> their
>>>>>>>> object into XML.  I do not have sufficient exposure to enough  
>>>>>>>> XML
>>>>>>>> libraries to know if this is truly an issue, however.
>>>>>>>
>>>>>>> I mentioned on the call that the people who would have problems
>>>>>>> with
>>>>>>> order
>>>>>>> are those using dumb or broken data binding tools. There are
>>>>>>> specs that
>>>>>>> have
>>>>>>> made and continue to make such allowances, but I think this is a
>>>>>>> lot of
>>>>>>> effort for the rest of us to help people using broken tools.
>>>>>>>
>>>>>>> (A similar argument exists regarding making allowances for
>>>>>>> people that
>>>>>>> insist on using non-XML tools to process XML.)
>>>>>>>
>>>>>>> -- Scott
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --------------------------------------------------------------------
>> -
>>>>>>> To unsubscribe from this mail list, you must leave the OASIS TC
>>>>>>> that
>>>>>>> generates this mail.  Follow this link to all your TCs in OASIS
>>>>>>> at:
>>>>>>> https://www.oasis-
>> open.org/apps/org/workgroup/portal/my_workgroups.php
>>>>>>
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe from this mail list, you must leave the OASIS TC
>>>>>> that
>>>>>> generates this mail.  Follow this link to all your TCs in OASIS  
>>>>>> at:
>>>>>> https://www.oasis-
>> open.org/apps/org/workgroup/portal/my_workgroups.php
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe from this mail list, you must leave the OASIS TC  
>>>>> that
>>>>> generates this mail.  Follow this link to all your TCs in OASIS  
>>>>> at:
>>>>> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>>>>>
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Gabe Wachob / gwachob@wachob.com \ http://beatslacker.com
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe from this mail list, you must leave the OASIS TC  
>>>> that
>>>> generates this mail.  Follow this link to all your TCs in OASIS at:
>>>> https://www.oasis-open.org/apps/org/workgroup/portal/
>>>> my_workgroups.php
>>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe from this mail list, you must leave the OASIS TC that
>>> generates this mail.  Follow this link to all your TCs in OASIS at:
>>> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe from this mail list, you must leave the OASIS TC that
>> generates this mail.  Follow this link to all your TCs in OASIS at:
>> https://www.oasis-open.org/apps/org/workgroup/portal/ 
>> my_workgroups.php
>
>