Right now we're extracting this information from Bitcoin Core with a clever algorithm that considers the returned blocks (there's a bug there btw: #15001), converts the result to s/b and rounds up to integer and asks estimatesmartfee between two upper and smaller values and results in a dictionary with the
- key: fee target
- value: integer satoshi/vbyte
This reduces the whole spectrum of 2-1008 feetargets to only a couple of entries today:
https://wasabiwallet.io/api/v2/btc/Blockchain/all-fees?estimateSmartFeeMode=CONSERVATIVE
https://wasabiwallet.io/api/v2/btc/Blockchain/all-fees?estimateSmartFeeMode=ECONOMICAL
Of course the worst case scenario we'd have to do 1007 estimatesmartfee query, but today our algorithm is able to establish this dictionary with only about 12.
With this we are able to show a more accurate pre-transaction build fee estimation than other wallets, who're blindly targeting one or two RPC targets and doesn't try to explore that if the same fee can apply to less targets, thus giving back less accurate feedback to the users, like the user targets 2 weeks, but the tx still confirms within 6 hours.
Anyway, it's no biggie, but it would be nice if we could reduce our RPC queries to only one, if Core would give back the "allfees" with just one request. Maybe an estimateallsmartfee CONSERVATIVE/ECONOMICAL RPC would be a good idea. Thoughts?
