[Laszlo-dev] datasets and XML namespace support

[Oliver Steele]

On Sep 12, 2005, at 3:11 PM, Henry Minsky wrote:
> OK, that's one vote for changing the default behavior. Do I hear  
> any others?

Yes, please.  But you have to phase it in so that it doesn't break  
compatibility.  One way to do this is to make the nsprefix flag  
optional in the next release; mandatory for non-SOLO apps in 3.2; and  
deprecate it in 3.3.  (This assumes there's a maintenance release  
before 3.2; otherwise increase all the numbers by 0.1.  It's also a  
somewhat accelerated deprecation schedule; if no one squawks, I'd  
still like to do this.)

> On 9/12/05, P T Withington <ptw at pobox.com> wrote:
>
>> It seems to me that the stripping of namespaces is a bug that should
>> be fixed.  If there are apps that depend on that (mis) feature, you
>> could have a deprecated compatibility flag for one release.
>>
>> On 12 Sep 2005, at 12:43, Henry Minsky wrote:
>>
>>
>>> I'm doing some design and implementation to add simple namespace
>>> support to the data APIs, and wanted to get some feedback from  
>>> people
>>> on how they think things would work best.
>>>
>>> I'm a little surprised more people haven't run into this, but I  
>>> guess
>>> SOLO deployment hasn't been out in the wild very long. The single  
>>> bug
>>> report I have gotten so far was from someone trying to use an RSS  
>>> feed
>>> in SOLO mode.
>>>
>>>
>>> Below is an excerpt from the bug I filed, and then some design
>>> proposals.
>>>
>>>
>>>  Key: LPP-653
>>>         URL: http://www.openlaszlo.org/jira/browse/LPP-653
>>>     Project: OpenLaszlo
>>>        Type: Bug
>>>  Components: Laszlo Foundation Classes - Data, Server - Data  
>>> Transport
>>>    Reporter: Henry Minsky
>>>    Priority: --
>>>
>>>
>>> consider this XML data
>>>
>>>
>>> <?xml version="1.0" encoding="UTF-8"?>
>>> <rdf:RDF
>>>  xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>>>  xmlns="http://purl.org/rss/1.0/"
>>>  xmlns:cc="http://web.resource.org/cc/"
>>>  xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/"
>>>  xmlns:dc="http://purl.org/dc/elements/1.1/"
>>>  xmlns:syn="http://purl.org/rss/1.0/modules/syndication/"
>>>  xmlns:admin="http://webns.net/mvcb/"
>>>
>>>
>>>>
>>>>
>>>>
>>>
>>>     <channel rdf:about="http://del.icio.us/gyardley">
>>>         <title>del.icio.us/gyardley</title>
>>>         <link>http://del.icio.us/gyardley</link>
>>>
>>>
>>> When loaded at runtime in SOLO mode, you get the XML basically as
>>> above, but proxied mode, the namespace prefixes are stripped off
>>> before the data is sent from the server, and the app sees
>>>
>>> <RDF>
>>>     <channel about=\"http://del.icio.us/gyardley\">
>>>             <title>del.icio.us/gyardley</title>
>>>             <link>http://del.icio.us/gyardley</link>
>>>
>>>
>>>
>>> Aside from general problem losing the namespace prefixes, the
>>> developer cannot
>>> use the same XPath queries in SOLO and proxied mode, because
>>> element names
>>> (and attribute names) are different; "rdf:RDF" vs. "RDF",  
>>> "rdf:about"
>>> vs. "about".
>>>
>>>
>>>
>>> + Proposal:
>>>
>>> [A] add a "nsprefix" flag to <dataset>:
>>>
>>>     <dataset name="lds" src="namespace.xml" nsprefix="false"
>>> request="true" type="http" />
>>>
>>> when nsprefix=false, namespace prefixes are discarded from  
>>> element and
>>> attribute names.
>>>
>>>
>>>     <dataset name="ldsns" src="namespace.xml" nsprefix="true"
>>> request="true" type="http" />
>>>
>>> when nsprefix=true, namespace prefixes are preserved on element and
>>> attribute names
>>>
>>>
>>> Currently, all data loads in SOLO mode act as if nsprefix=true,  
>>> due to
>>> the Flash XML parser behavior, and all data loads in proxied mode  
>>> acts
>>> as if nsprefix=false, because we've always done it that way.
>>>
>>>
>>> [B] I propose not supporting nsprefix=false in SOLO mode, for the
>>> time being
>>>
>>> It would be easy to write the string string frobbing code in the LFC
>>> to strip namespace prefixes from data nodes, in order to make it act
>>> like proxied data does now, but it seems to me that would be a
>>> performance hit that not many developers would want to choose, given
>>> how slow actionscript is already.
>>>
>>> [C] What should be done about compile-time inline data?
>>>
>>> If datasets are specified as filenames, the data is compiled into  
>>> the
>>> app. In that case we have full control over what it looks like, in
>>> either SOLO or proxied mode:
>>>
>>>     <dataset name="localds" src="namespace.xml" nsprefix="false" />
>>>
>>>     <dataset name="localdsns" src="namespace.xml" nsprefix="true" />
>>>
>>> The question is what should the default setting of the nsprefix
>>> flag be?
>>>
>>> Should it remain false in both the SOLO and proxied case? Or  
>>> should it
>>> default to true in the SOLO case, so that these cases return
>>> exactly the
>>> same values for node and attribute names in SOLO mode?
>>>
>>>
>>>     <dataset name="foo" src="namespace.xml" />         # inline
>>>
>>>     <dataset name="bar" src="namespace.xml"  type="http" /> # http
>>> runtime loaded
>>>
>>>
>>>
>>>
>>> --
>>> Henry Minsky
>>> Software Architect
>>> hminsky at laszlosystems.com
>>>
>>> _______________________________________________
>>> Laszlo-dev mailing list
>>> Laszlo-dev at openlaszlo.org
>>> http://www.openlaszlo.org/mailman/listinfo/laszlo-dev
>>>
>>>
>>
>>
>>
>
>
> -- 
> Henry Minsky
> Software Architect
> hminsky at laszlosystems.com
>
> _______________________________________________
> Laszlo-dev mailing list
> Laszlo-dev at openlaszlo.org
> http://www.openlaszlo.org/mailman/listinfo/laszlo-dev
>