Greg Wilkins

@meltsufin I think the image should work out of the box on all reasonable containers. Currently I can't even build on a machine with 32GB because the tests fail because...

@joakime I'm not sure we have time to do the resource service this week, but I think you and I should work on this - perhaps I'll do the URI...

Note the other thing we need to think about with regards to resource service, is should the mimetype stuff move from the servlet contexts to the core context - ie...

@joakime I'm not exactly on the same page as you... maybe it is naming. Firstly I don't think we need new Request methods, rather we need to better define the...

Interestingly enough we already have `String URIUtil.getCanonicalPath(String)` and `String URIUtil.getDecodedPath(String)`, with the former doing just dot segment removal and the later doing full % decoding. They are called moderately frequently,...

@joakime I don't think you have interpreted the RFC correctly. Specifically `file:/tmp/foo%2cbar` and `file:/tmp/foo,bar` are indeed the same resource. From [RFC 3986 2.4 When to Encode or Decode]( https://datatracker.ietf.org/doc/html/rfc3986#section-2.4): >...

> These are not equivalent, and don't point to the same resource. > > * https://www.w3.org/Addressing/URL/uri-spec.html > * https://www.w3.org/Addressing/URL%2Furi-spec.html Exactly! `%2F` is meaningful in both `http:` and `file:` schemes. But...

@sbordet that doesn't need to be encoded for the sake of interpreting the URI. It is not ambiguous in a URI path as `\` is not significant. It is significant...

The pathInContext is encoded... but only canonically encoded, i.e. it encodes only characters like %2F. I expect the problem is that URI is even more picky about what characters and...

Thanks, Good catch. However, I'm wondering about if we should always start the swapped handler? Perhaps we start it only if the handler we are replacing was started?