-
@ mleku
2025-06-07 08:54:44
i'm gonna just reintegrate the code i already have written for correctly handling this, adjusting it to work with the "normal" string fields that i reverted to using.
the reason being that i already know that the old realy json codec was the damned fastest one ever written and in fact it will be come a little faster because the hex fields, id, pubkey and sig are all not any more going to be de/coded. so, a bit more memory management but probably shaves a few microseconds off this step
already almost completed the work getting the encode side refactored, just need to finish on the tags. they are now slices of slices of strings so actually i might just shortcut this and use stdlib json encode for this step. but no, the escaping. so i have to rewrite the tag marshaling step as well.