Hi Peter, at the moment I have no real patches to send, we just have a small bug fix for the SPARQL-JSON writer that we currently host locally in our source code, but would replace with your new version once it is available.
But if you want to I can rewrite your JSON result parsers so they use Jackson - should be fairly straightforward. Your result writers seem to directly write to the output streams anyways. If I find the time I will create a pull request in your bitbucket repository :-) Jackson is, BTW, FMPOV the right decision. It is appropriately licensed and by far the fastest Java JSON writer and parser around. It is even much faster than Java's Object Streams for marshalling data. In the coming months, I intend to work a bit on punctual improvements of SPARQL performance also. Maybe there will be some contributions there as well. :) Greetings, Sebastian 2013/3/28 Peter Ansell <ansell.pe...@gmail.com> > Hi Sebastian, > > I implemented the SPARQL JSON reader since beta2 and that > implementation uses json.org. Since then I implemented an RDF/JSON > parser using Jackson and based on that experience I am keen to switch > to using it in Sesame for both reading and writing of SPARQL Results > JSON, but haven't yet found the chance to implement it. (Feel free to > donate code changes to Sesame if you have patched versions already > using Jackson internally!) > > The org.json code is also used in the Sesame workbench module, which > isn't likely to be a dependency of Marmotta. > > Peter > > On 27 March 2013 19:15, Sebastian Schaffert <sschaff...@apache.org> wrote: > > Hi Peter, > > > > funny, we don't have any dependency to the json.org module left in > > Marmotta. Where is Sesame actually using it? The SPARQL JSON writer seems > > to write directly to the output stream. > > > > Maybe the beta2 does not yet include this code anyways? If you have a new > > release with more SPARQL JSON support, we would be very happy, because > then > > we can throw away our own code :) > > > > :-) > > > > 2013/3/27 Peter Ansell <ansell.pe...@gmail.com> > > > >> On 26 March 2013 23:04, Andy Seaborne <a...@apache.org> wrote: > >> > On 26/03/13 12:47, Sergio Fernández wrote: > >> >> > >> >> Hi, > >> >> > >> >> On 26/03/13 12:33, Andy Seaborne wrote: > >> >>> > >> >>> On 25/03/13 16:50, Sergio Fernández wrote: > >> >>>> > >> >>>> On 25/03/13 17:02, Andy Seaborne wrote: > >> >>>>> > >> >>>>> The license for json.org has the linethat makes it in > contradiction > >> >>>>> with the definition of open source: > >> >>>> > >> >>>> > >> >>>> This is no longer an issue, since we managed to rewrite some code > and > >> >>>> completely removed such dependency (commit > >> >>>> bdb25fedcbb38c311ffb4f08e84505163109718d) with other with better > >> >>>> licensing and performance. > >> >>> > >> >>> > >> >>> I would point out that by shipping any system, you take on the > >> >>> responsibilities from the license of that system system, not the > part > >> >>> you happen to use. > >> >> > >> >> > >> >> Sure. In this case the system is Jackson, which was already in use in > >> >> Marmotta: I've just replaced the single usage of json.org by the > >> >> equivalent in Jackson. > >> >> > >> >> Thanks for the remark; we really appreciate your toil as mentors :-) > >> > > >> > > >> > In this case, having any system like Sesame mention in their NOTICE > file, > >> > recursively brings it in. json.org is particular fiddly as so many > >> systems > >> > include it by source (and forget to be explicit about it). > >> > >> I have it on my todo list to replace org.json with Jackson in Sesame. > >> I wrote a Talis RDF/JSON parser and writer using Jackson the other day > >> and it works famously, so it should be fairly simple to convert the > >> Sesame Workbench and the SPARQL Results JSON module when I have time. > >> > >> Cheers, > >> > >> Peter > >> >
