[
https://issues.apache.org/jira/browse/JENA-162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13149940#comment-13149940
]
Andy Seaborne commented on JENA-162:
------------------------------------
The dataset description is still parsed out of the protocol by Fuseki (Fuseki
being the SPARQL protocol implementation). It then places it in the execution
environment. There's now an (ARQ) class DatasetDescription to hold the two
lists of URIs (strings). My style (these days) is to name things that are
architectural concepts.
Various details get put in the context of a QueryExecution - the query, the
algebra the value of now() and more.
You can then access the dataset description object from the protocol and the
dataset description in the query at any point.
The only downside is that the rule that the protocol dataset description
modifies the request in preference to the query isn't enforced. There is all
information around and having the precedence rule in a static method seems
adequate. See (in ARQ) DatasetDescription.create(Query, Context).
When it gets to the query engine, no decisions have been made - a convenience
piece of code captures the SPARQL decision rule.
> Add FROM/FROM NAMED support to Fuseki
> -------------------------------------
>
> Key: JENA-162
> URL: https://issues.apache.org/jira/browse/JENA-162
> Project: Jena
> Issue Type: New Feature
> Components: Fuseki
> Affects Versions: Fuseki 0.2.1
> Reporter: Rob Vesse
> Labels: from, from-named, sparql
> Attachments: FusekiFromSupport.patch
>
>
> Fuseki currently does not support FROM/FROM NAMED clauses returning a HTTP
> 400 error if a user tries to make use of them
> The attached patch provides for optional enabling of this feature by a user
> specified --from or --allowFrom switch when the server is started or
> programmtically via ARQ.getContext().set(Fuseki.enableFromClauses, true).
> Therefore unless a user turns this feature on by default the old behavior
> remains in place.
> When enabled Fuseki will not reject queries with FROM/FROM NAMED in them and
> will use DynamicDatasets.dynamicDataset() to build the dataset for queries,
> there is minimal overhead for queries without FROM/FROM NAMED clauses as the
> code in DynamicDatasets appears to leave the default dataset unchanged if no
> dataset description is present in the query AFAICT
> A small number of tests has been added to TestQuery.java to test the
> behaviour, more tests might be desirable and I can probably supply those if
> necessary but the ones given test all the obvious combinations with the
> trivial test data used by TestQuery
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
