[
https://issues.apache.org/jira/browse/JENA-2316?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andy Seaborne resolved JENA-2316.
---------------------------------
Resolution: Fixed
> SHACL: SPARQL validation message formatting inconsistency
> ---------------------------------------------------------
>
> Key: JENA-2316
> URL: https://issues.apache.org/jira/browse/JENA-2316
> Project: Apache Jena
> Issue Type: Bug
> Components: SHACL
> Reporter: Jakub Wach
> Assignee: Andy Seaborne
> Priority: Minor
> Fix For: Jena 4.5.0
>
> Attachments:
> SHACL__SPARQL_validation_message_formatting_inconsistency.patch
>
>
> It seems that SparqlValidation class inconsistently formats the nodes in the
> messages while reporting the issues:
> * default message "SPARQL SELECT constraint for" uses
> ShLib.displayStr(valueNode)
> * custom message provided in _sh:message_ is using NodeFmtLib.str
> The issues:
> * Inconsistency
> ** On the default settings there is no difference.
> ** However, ShLib formats nodes (at least references) according to the
> ShLib.nodeFmtAbbrev (it is not final).
> ** Two separate formatter instances could present different and inconsistent
> results.
> * Why static non-configurable formatting of nodes might be a bad idea?
> ** Apart from possibility to formatting the literals (dates) in a
> controllable and configurable way...
> ** Bnode labels are also affected. Default (NodeToLabel) formatting is using
> X00/X00X000 escaping (by default this is the way how bnode labels are
> presented in those messages). This escaping is non-deterministic (unlike
> UTF8).
> ** Ability in ShLib to modify the NodeFormatterTTL addressed those issues
> (although static configuration has its own issues) by making it at least
> controllable (e.g. by using NodeToLabel.createBNodeByLabelAsGiven).
> Path for that possibly very simple fix is provided.
> Way to reproduce issue: JENA-2315 (just comment sh:message in shape
> definition)
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
