mleku on Nostr: relayer breaks apart the interfaces and composes them, and that lets you extract them ...

relayer breaks apart the interfaces and composes them, and that lets you extract them if they exist to do interface specific things

https://realy.lol contains primarily libraries, the codebase parts of the actual relay are trivial compared to the library parts

the big long top level directory is that way because it makes importing things simpler, and lays out all the parts with human readable names from top to bottom. it's not quite idiomatic but when you see how much it cleans up your import blocks and makes them understandable you will agree

relayer, also, part of why i preferred it, is better at managing the number of threads in the websocket listeners, i had a huge problem with khatru's socket design that made it unsuitable for production in even... i mean, can you imagine? i don't even understand how khatru based relays are still functioning the excessive amount of threading they use

also, yeah, not to forget, i totally flattened the concurrency of the event store, so, it also is a solution for that, excessive amounts of threads in each query were a component of the resource leak problem i found with khatru/eventstore design

i have also added, btw, an index to the event store that lets you go from event serial to having the pubkey, event ID and created_at of the event, without having to decode any JSON, this was built to enable the different API i made that the /filter endpoint only returns event IDs, meaning the client can then paginate them arbitrarily, without pushing that work to the relay

there are some cases where it's better to paginate on the networked DB and i don't think that the client use case is that, this way, clients can grab a whole massive list of event IDs matching a query and pick at them as the user wants to see them