Erick λ
Erick λ
> Thanks! There is indeed a bug as we do not check for invalid characters properly, but `decode()` still fails later when it tries to verify checksums, with an ugly...
> Rebase conflict is preventing CI from running for some reason... Rebased
> Needs to update `buildroute` as well so that we do not allow building a route without a payment_secret: > > https://github.com/lightningnetwork/lnd/blob/fe405426cac5096773cc039dff1cb20367aeb54d/lnrpc/routerrpc/router_server.go#L1506-L1511 Does keysend use the `BuildRoute` function? If so,...
Yes, I agree. Since keysend relies on invoices with zero payment addresses, phase 2 might be problematic. > 2. Reject decoding invoices that don't have a payment secret. Maybe we...
cc @brunoerg
> > LND only processes version 0 SegWit programs (checking for exactly 20 or 32 bytes) and silently ignores versions 1-16, including Taproot (v1) addresses. Furthermore, once it finds a...
> [basemodule.cpp](https://github.com/bitcoinfuzz/bitcoinfuzz/blob/v2/include/bitcoinfuzz/basemodule.cpp) is where would would supply new fuzzing entry points for the BIP32 keys, e.g., extended_pub_key_deserialize, etc. I can imagine there being several such functions - some specific to...
**Note: I am not a mantainer to determine what can be added or not, but these are my thoughts:** > * Currently, I don't see support for static test vectors....
Nice! Thanks for starting this! :) > How would you like us to structure the PRs? One PR per (target, module) pair or implement at least two modules at once,...
I made a simple test to demonstrate the parsing discrepancy: ```rust #[test] fn test_raw_data_base32_roundtrip() { use bech32::Fe32; // These are the expected Fe32 values that should round-trip correctly // The...