We currently configure the ugorji library to be very loose when decoding. For example, spans was defined as a [][]string, but when the underlying message format changed there was no error because ugorji silently did the conversion.
Should we continue to do this? Do we want XMiDT cluster code to throw an error when a msgpack or JSON message is encountered that does not strictly conform to the WRP spec?
We currently configure the ugorji library to be very loose when decoding. For example,
spanswas defined as a[][]string, but when the underlying message format changed there was no error because ugorji silently did the conversion.Should we continue to do this? Do we want XMiDT cluster code to throw an error when a msgpack or JSON message is encountered that does not strictly conform to the WRP spec?