Hello, I am developing multi coin miner, and I would like to add decred support. I have some problems figuring block header assembly. So far I used official block header specification (https://docs.decred.org/advanced/block-header-specifications/) , and gominer source code to find out how to assemble block. I used stratum+tcp://yiimp.ccminer.org:4252 as job source. So far I figured out how to assemple most of the block header from stratum data. Version in LE(sent in job notification from stratum ) = 4 bytes, prevhash as int in BE, I need to reverse each int to LittleEndian, coinbase1 - 112 bytes, assembled by pool, contains everything up to "Extra Data" from header specification, and also contains 8 bytes of this extra data. Now, I have 32 bytes left (ExtraData - 8 bytes, specified in coinbase1). I presume this data is up to client to make up. part of this data is probably extranonce1 + extranonce2. Stratum subscription responds with 12 bytes of extranonce1 and extranonce2_len of 12. so that is 24 bytes, and I have 8 bytes left. This 8 bytes is the problem , I am not sure what it is. Is it my nonce space ? Also, how do extranonce1, extranonce2 fit into block header ? My current idea is: Block Header : version from job : 4 bytes in LE, stays in LE, prevhash : 32 bytes, needs to be byteswapped per 4 bytes, coinbase1: 112 bytes, from pool, no modification, extranonce1: 12 bytes, from pool, no modification, extranonce2: 12 bytes, length specified by pool, generated by miner, sent as little-endian hex string. nonce : 8 bytes - rolling counter, sent in job notification as little-endian hexadecimal string. Block header is then hashed, I currently have no idea about endiannes of this has, and whether last N bytes or First N bytes are checked against target. Any miner developers here willing to help ? thank you.