model

This hierarchy is meant to reflect the output of com.rabbitmq.client.impl.ValueReader.readFieldValue in a type-safe way.

Note that we don't include LongString here because of some ambiguity in how RabbitMQ's Java client deals with it. While it will happily write out LongStrings and Strings separately, when reading it will always interpret a String as a LongString and so will never return a normal String. This means that if we included separate LongStringVal and StringVals we could have encode-decode round-trip differences (taking a String sending it off to RabbitMQ and reading it back will result in a LongString). We therefore collapse both LongStrings and Strings into a single StringVal backed by an ordinary String.

Note that this type hierarchy is NOT exactly identical to the AMQP 0-9-1 spec. This is partially because RabbitMQ does not exactly follow the spec itself (see https://www.rabbitmq.com/amqp-0-9-1-errata.html#section_3) and also because the underlying Java client chooses to try to map the RabbitMQ types into Java primitive types when possible, collapsing a lot of the signed and unsigned types because Java does not have the signed and unsigned equivalents.

A string whose UTF-8 encoded representation is 255 bytes or less.

Parts of the AMQP spec call for the use of such strings.