From de672239befaa9a79eb7836a4b3f8521db670d46 Mon Sep 17 00:00:00 2001 From: 20kdc Date: Wed, 16 Aug 2017 15:37:33 +0100 Subject: [PATCH] Update protocol.0 --- protocol.0 | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/protocol.0 b/protocol.0 index fcd5198..94332fc 100644 --- a/protocol.0 +++ b/protocol.0 @@ -22,8 +22,10 @@ A name (as a length-minus-1-byte-prefixed-string), another name (in the same format), and the rest is data. -Copper packet data may be up to 1506 bytes long - - this does not include header data, which may be up to (256*2) + 3 bytes long. +Copper packet data may be up to 1507 bytes long. + 1500 in-relib data bytes + 7 relib bytes. +This makes the user's life easier by having an easy to remember MTU from within relib of 1500 bytes. +This does not include header data, which may be up to (256*2) + 3 bytes long. Loop detection should performed by checking if a packet exactly the same has been seen recently - other rejection, alteration and routing measures @@ -84,3 +86,8 @@ The connection between OpenComputers and Copper can be performed with a TCP gate It doesn't really matter how the data is transmitted here, since it's just a gateway. Possibly consider integrating a hierarchial gateway into the OpenComputers/world bridge. + +ERRATA: + +The data MTU was described as 1506 bytes due to an error in the protocol.2 document. +It should be 1507 to fit 1500 relib byets and the 7-byte header.