The question seems anectodic but it has an importance when reading the old key.MakeNewKey() source code :
- At that time, coinbase transactions of 50Btc were sent to a newly generated pubkey at each blocks. As far I can tell, each key were added to the wallet at the approximately the block’s timestamp.
- Bitcoin core was using
/dev/urandomon Linux but on Windows even when openssl was asked for secure randomness, the generator had to be initialized manually each time to a initial value : and in the case of key generation, that value was set to HKEY_PERFORMANCE_DATA by Satoshi !
So if I’m not wrong, this means coinbases mined through using Windows had weak randomness as the possible values are far more limited than the secp256k1’s order. As a result, does this means he only used Linux personally ?
