Pierre Krieger

The current situation is that Beefy is no longer recognized by smoldot. Instead, beefy log items are treated as "unknown". We have interest in supporting Beefy because it can provide...

This feature conflicts with warp syncing. It is possible to warp sync over the block specified in a `codeSubstitutes`, and not be aware that a substitution needs to take place.

> This feature conflicts with warp syncing. It is possible to warp sync over the block specified in a codeSubstitutes, and not be aware that a substitution needs to take...

This is not as important as it looks, as right now we don't keep track of the announced blocks except to know which node knows which block.

https://github.com/paritytech/smoldot/pull/2631 doesn't solve the issue but changes the panic message to include the service that has a bug in it, so if it happens again it would be nice to...

Just personally encountered this: > panicked at 'block already unpinned for subscribe-best subscription', /home/tomaka/Projets/smoldot/bin/light-base/src/runtime_service.rs:394:25 It's unclear whether it's the same panic as the OP, but I'm going to go with...

The subscription code is definitely behaving as intended. There's probably a bug in the runtime service unfortunately.

This newly-added debug_assertion panics: https://github.com/paritytech/smoldot/blob/c537c408c4382eab79d56029bc6c24f6a9ab9f6f/bin/light-base/src/runtime_service.rs#L317

Problem comes from here: https://github.com/paritytech/substrate-lite/blob/ca2bc4a01e1ef52a81595a80c7f064c2ebaf0500/bin/light-base/src/sync_service/parachain.rs#L571-L574 We report blocks in a completely incorrect way.

Linking here for reference. Encrypted memory: https://docs.sequoia-pgp.org/sequoia_openpgp/crypto/mem/struct.Encrypted.html https://docs.sequoia-pgp.org/src/sequoia_openpgp/crypto/mem.rs.html#217-220