Could RDFParserRegistry::getRegistered and ResultSetReaderRegistry::getRegistered be added?
On Fri, May 20, 2022 at 9:01 PM Andy Seaborne <a...@apache.org> wrote: > > > > On 20/05/2022 14:05, Martynas Jusevičius wrote: > > Andy, is that correct? > > Yes > > Andy > > > > > On Tue, May 17, 2022 at 1:33 PM Martynas Jusevičius > > <marty...@atomgraph.com> wrote: > >> > >> On Tue, May 17, 2022 at 1:19 PM Andy Seaborne <a...@apache.org> wrote: > >>> > >>> RDFLanguages is a general registry of names (Lang's) in the system. > >>> > >>> It is not for functionality. > >>> > >>> RDFParserRegistry > >>> RDFWriterRegistry > >>> RowSetReaderRegistry, ResultSetReaderRegistry > >>> RowSetWriterRegistry, ResultSetWriterRegistry > >>> StreamRDFWriter > >>> > >>> A Lang needs looking up in a registry to see if there is support for it. > >> > >> Thanks, I didn't know these existed. > >> > >> But there are no RDFParserRegistry::getRegistered or > >> ResultSetReaderRegistry::getRegistered methods? > >> > >> So do I still need to iterate RDFLanguages::getRegistered and check > >> each Lang against > >> RDFParserRegistry::isRegistered/ResultSetReaderRegistry::isRegistered? > >> > >>> > >>> Andy > >>> > >>> On 17/05/2022 09:54, Martynas Jusevičius wrote: > >>>> Hi, > >>>> > >>>> After upgrading from 4.3.2 to 4.5.0, some of our RDF writing code > >>>> started failing. > >>>> > >>>> It seems that this is due to RDFLanguages.isTriples(Lang.SHACLC) > >>>> returning true, which messes up our content negotiation as it attempts > >>>> to write Models as SHACLC. Can this be rectified? > >>>> > >>>> The RDFLanguages registry is a bit of an oxymoron in general. Right > >>>> now it's a bag of all sorts of syntaxes Jena supports, half of which > >>>> are not even "RDF languages". We need to iterate and filter the > >>>> languages just to know which ones can be used to read/write Models, > >>>> which can be used for ResultSets etc.: > >>>> https://github.com/AtomGraph/Core/blob/master/src/main/java/com/atomgraph/core/MediaTypes.java#L86 > >>>> Wouldn't it make sense to have separate registries depending on the > >>>> entity types they apply to? > >>>> > >>>> Thanks. > >>>> > >>>> Martynas
