That's exactly correct. Dave
-----Original Message----- From: Andy Seaborne [mailto:[email protected]] Sent: Thursday, May 02, 2013 5:00 PM To: [email protected] Subject: Re: Jena 2.10.1 / SDB 1.3.6 -- testing for next release On 01/05/13 20:11, Andy Seaborne wrote: > Dave, > > I can't recreate this. When you say "does not work" do you mean an > exception? Or parse error? I can see how it might happen. Are you getting an exception from IRIResolver? About an UNWISE_CHARACTER? Recorded as JENA-448 Andy > > I tried this: > > public static void main(String... args) > { > String fn = "C:\\home\\afs\\D.ttl" ; > Model m = ModelFactory.createDefaultModel() ; > FileManager.get().readModel(m, fn) ; > m.write(System.out, "TTL") ; > System.out.println("DONE") ; > } > > with a classpath that include jena-arq, jena-core, etc. > and variations with / and they all seemed to work. > > Do you have a complete, standalone example I can try out? > > Andy > > On 01/05/13 16:00, Lebling, David (US SSA) wrote: >> I retried all this just to be certain. >> >> FileManager.readModel(model, "c:/X/base/sub/file.rdf") >> >> i.e., the "cleaned up" filename, does work. The "unclean" >> ("C:\X\.....") version does not. >> >> Apologies for my confusion. Previously I didn't have to clean the >> filename for it to work; that part of my original message is correct. >> I -have- always had to clean the filename of the store descriptor >> file for SDBFactory.connectStore(String file). >> >> In the process I also retested with transaction isolation set higher >> (to Connection.TRANSACTION_REPEATABLE_READ), and the duplicate node >> error (JENA-447) still happens. >> >> I do not use RDFDataMgr (unless something uses it indirectly). >> >> Dave >> >> -----Original Message----- >> From: Andy Seaborne [mailto:[email protected]] >> Sent: Wednesday, May 01, 2013 10:00 AM >> To: [email protected] >> Subject: Re: Jena 2.10.1 / SDB 1.3.6 -- testing for next release >> >> On 01/05/13 14:41, Lebling, David (US SSA) wrote: >>> Sorry, I mistyped. It's >>> >>> FileManager.readModel(Model model, String filenameOrURI) >>> >>> The offending filenames were all Windows style, in fact mixed >>> slashes and backslashes, such as: >>> >>> C:\X\base\sub/file.rdf >>> >>> That used to work. Even "cleaning" the filename to: >>> >>> c:/X/base/sub/file.rdf >>> >>> didn't work. >>> >>> I hope this helps. >> >> Yes - it gives be something to look it. >> >> The code has to special-case C: ... because technically it is a URI, >> with schema name "C" (oops). >> >> Does RDFDataMgr work? >> >> Andy >> >> >>> >>> Dave >>> >>> -----Original Message----- >>> From: Andy Seaborne [mailto:[email protected]] >>> Sent: Wednesday, May 01, 2013 9:01 AM >>> To: [email protected] >>> Subject: Re: Jena 2.10.1 / SDB 1.3.6 -- testing for next release >>> >>> On 30/04/13 22:10, Lebling, David (US SSA) wrote: >>>> Andy, >>>> >>>> I was able to get some time today and managed to install and run >>>> with jena-sdb-1.3.6-SNAPSHOT on top of apache-jena-2.10.1-SNAPSHOT. >>>> >>>> I ran a bunch of our integration tests and one of our stress tests >>>> (not the one that uses the transaction isolation code, yet -- maybe >>>> tomorrow). All of them ran fine. >>>> >>>> The only potential issue I found was that >>>> FileManager.readModel(String uri, String filenameOrUri) now seems >>>> to actually require that the second argument be a URI. I was >>>> running on Windows and before calling the IRI code FileManager >>>> apparently isn't detecting and converting file names to URIs. It >>>> may also be that the Javadoc hasn't caught up with a change to the >>>> readModel contract, if it now requires a URI. I just forced a URI to work >>>> around the issue. >>>> >>>> Dave >>>> >>> >>> Hi Dave, >>> >>> Thanks for the report. >>> >>> Which FileManager operation did you mean ? >>> (there isn't a readModel(String, String)) >>> >>> FileManager.loadModel(String filenameOrURI, String rdfSyntax) >>> FileManager.readModel(Model model, String filenameOrURI) >>> FileManager.readModel(Model model, String filenameOrURI, String >>> rdfSyntax) >>> >>> and they seem to work for me. >>> >>> Andy >>> >> >
