Am 11.12.2011 um 18:08 schrieb Damian Steer: > > On 9 Dec 2011, at 16:25, Thorsten Möller wrote: >>> >>>> Which brings me to ask/propose the following: >>>> Would it be possible to check whether the functionality used from >>>> ICU4J is provided similarly meanwhile by the JDK (or another library >>>> that is already a dependency)? If possible - if its possible to >>>> implement it in an alternative way - then this (heavyweight) >>>> dependency can be removed completely. >>> >>> That would be good - could you have a look? >> >> After checking in more detail, I got to a point where most was replaced by >> built-in JDK classes. However, there are some references left in >> com.hp.hpl.jena.iri.impl.AbsLexer and one in >> com.hp.hpl.jena.rdf.arp.impl.CharacterModel which I don't know how to >> resolve due to lack of Unicode experiences nor enough time resources that I >> can spend on that. If you want I can create patches and send them to you - >> unfinished work, though. I also tried to simply comment the remaining >> references (they all seem to be kind of syntax checks) and run the JUnit >> tests, but that apparently failed. >> >> Thorsten > > Hi Thorsten, > > This is great (begone megabytes of dependencies!). > > Could you open a ticket and upload the patch? I'll have a crack at completing > this work, if possible.
Done (JENA-177). > Which references were giving you problems? com.hp.hpl.jena.rdf.arp.impl.CharacterModel, line 147: - reference to UCharacter.getCombiningClass(..) com.hp.hpl.jena.iri.impl.AbsLexer: - Normalizer.quickCheck(..) not available in java.text.Normalizer - some references to UCharacter and UProperty These references seem to need higher level changes in order to resolve them, hopefully. HTH, Thorsten
