Jonas Konrad

closing as fixed by https://github.com/FasterXML/jackson-databind/issues/3102

You can generate a key pair with this snippet: ```java KeyPairGenerator keyGen = KeyPairGenerator.getInstance("RSA"); keyGen.initialize(2048); KeyPair pair = keyGen.generateKeyPair(); keyProvider = new KeyPairWrapper(pair); Buffer.PlainBuffer buf = new Buffer.PlainBuffer().putPublicKey(pair.getPublic()); publicKey =...

not sure we can change the existing behavior, but it seems reasonable to add an api to get the unencoded path component.

the test passes for me on linux, without redis running.

annoying. it needs to stay in onRequestPipelineBuilt for http2 but has to go into onStreamPipelineBuilt for http1. and it may break entirely in a future update where i replace pipeliningserverhandler...

ive merged 4.2.x and now the logbook test suite tests fail

@dstepanov i assigned you since apparently you want to fix propagation :smile: this pr can wait until then.

The `ContentNegotiationSpec` still fails because of a test case that has `Accept: application/json, application/xml`, and expects that to only match the controller for `application/json`. So basically it relies on the...

Apparently others have noticed the Edge thing too, and they come to the same conclusion: https://github.com/willdurand/Negotiation/issues/98#issuecomment-390625746 They also propose the same approach of prioritizing non-wildcard accept types that I use...

yea i think it's not worth a config option. I'll let the PR sit and we can revisit it for 4.0