clj-protobuf.core — com.github.bpalermo/clj-protobuf 0.1.5

Serialization, on top of what the generated code produces.

Generated X->proto / proto->X fns are Message-in/Message-out — they never touch bytes. This namespace is the other half: Message to bytes and back.

Typical round trip:

(-> rec HelloRequest->proto pb/encode)                     ; -> bytes
(->> bytes (pb/decode HelloRequest-prototype) proto->HelloRequest)

decode takes the generated X-prototype var — the same value everything else in the contract keys off — so there is exactly one handle per message type in user code.

decode clj

(decode prototype input)

Parse one message. prototype is the generated X-prototype var; input is a byte array, ByteString, ByteBuffer, or InputStream (read to the end).

decode-delimited clj

(decode-delimited prototype in)

Parse one length-prefixed message from a stream; nil at end of stream.

encode clj

(encode msg)

(encode msg out)

Message -> bytes. With an OutputStream, writes to it and returns it.

encode-delimited clj

(encode-delimited msg out)

Write msg length-prefixed, for framed streams of messages.

unknown-fields clj

(unknown-fields msg)

The unknown fields protobuf-java preserved on a parsed Message. Preserved on the Message; necessarily dropped by a record round trip (proto->X reads declared fields only), so re-encoding a record is not byte-identical to the original when unknown fields were present. Inspect them here when that matters.