Re: CMS diff: Jena Full Text Search

2019-01-29 Thread Andy Seaborne 




On 28/01/2019 21:01, vincent ventresque wrote:

Hi Ajs6f

Thanks for including me in the conversation, but I have to confess I've 
never looked at java classes (I only use command line tools).


Le 28/01/2019 à 21:05, ajs6f a écrit :
On Jan 28, 2019, at 2:57 PM, Chris Tomlinson 
 wrote:


Hi Adam,

I haven’t seen that error. What I’ve done in the past is to replace 
the jena-text doc file with the new contents in Eclipse in an SVN 
checkout of the jena-doc-site and then committed.
I can definitely do that (and will when we're happy with the patch), 
but see below.



Out of curiosity when is it necessary to use the

 [] ja:loadClass "org.apache.jena.tdb.TDB” .

and

    [] ja:loadClass   "org.apache.jena.query.text.TextQuery” .


It's not necessary any more.

The ServiceLoader Jena initialization does this.

The other initialization step should also be unnecessary:

 tdb:DatasetTDB  rdfs:subClassOf  ja:RDFDataset .
 tdb:GraphTDBrdfs:subClassOf  ja:Model .



? I do not use them in the config when running fuseki war in tomcat.
I have no idea whatsoever! :grin: I wouldn't have thought them needed 
either.


Vincent-- any comment?

ajs6f



Regards,
Chris




On Jan 28, 2019, at 11:11 AM, ajs6f  wrote:

Recently Vincent offered a nice patch to our text indexing 
documentation, as shown below. Oddly, when I now go to merge it (a 
bit late, sorry!), I get an error: "Can't locate anonymous's tree to 
clone". Is anyone familiar with that? I know very little about the 
SVN-based CMS, so I'm not even sure where to start looking...


ajs6f

Initialization

2019-01-29 Thread Andy Seaborne 

Chris wrote (on users@):

>>> Out of curiosity when is it necessary to use the
>>>
>>>  [] ja:loadClass "org.apache.jena.tdb.TDB” .
>>>
>>> and
>>>
>>> [] ja:loadClass   "org.apache.jena.query.text.TextQuery” .


It's not necessary any more.

The ServiceLoader Jena initialization does this.

(PS checking - maybe jena-permissions needs its loadClass - I don't see 
a service file)


Fuseki doc says:
  # Load custom code (rarely needed)
  # ja:loadClass "your.code.Class" ;

Only local customizations would need "loadClass".

Any mentions in any of the Jena documentation can be removed.

I've just gone though the website markup and removed all mentions in 
assemblers.



The other initialization step should also be unnecessary:

 tdb:DatasetTDB  rdfs:subClassOf  ja:RDFDataset .
 tdb:GraphTDBrdfs:subClassOf  ja:Model .

is now done in init java code (VocabTDB):

   AssemblerUtils.registerDataset(tDatasetTDB,
  new DatasetAssemblerTDB());
   AssemblerUtils.registerModel(tGraphTDB,
  new TDBGraphAssembler());

(SDB needs it still)

Andy