When i checked - and there was a comment on that - we were ok relyibg on should and not must. Ijson was not a great enforcement from memory.
Le lun. 14 mai 2018 08:23, Mark Struberg <[email protected]> a écrit : > Hi folks! > > I got a mail from Anders Rundgren (OpenAPI) which points out a spec glitch > with JSON-B > > It's basically whether the spec correctly forces the I-JSON option. > javax.json.bind.JsonbConfig#withStrictIJSON > > The spec is really thin in this topic. I'd personally interpret it as: we > should be I-JSON conform. > > You can read up more about I-JSON on the JSON spec > https://tools.ietf.org/html/rfc7159 > just grep for 'interop'. > As far as I understand I-JSON is standard JSON, but avoiding all the > possible portability traps, right? > > While the problems with Yasson have nothing to do with us, I'd rather love > to check whether we also have the same bug in Apache Johnzon. > And then probably provide a unit test we can apply to Johnzon but probably > even submit to the TCK (under ALv2 ofc). > > LieGrue, > strub > >
