Fix/ln dlc/renew after disconnect #143
Conversation
Tibo-lg
force-pushed
the
fix/ln-dlc/renew-after-disconnect
branch
3 times, most recently
from
456a63e to
a8070aa
Compare
|
@luckysori I think this is finally ready to be reviewed. There was actually several bugs related to revocation secrets. I added checks that helped find the issues, but I also opened #146 as I think it'd be better to have tests for that. |
| @@ -60,7 +60,7 @@ impl_dlc_writeable_enum!( | |||
| (6, RenewOffered, {(offered_contract_id, writeable), (counter_payout, writeable), (is_offer, writeable), (offer_next_per_update_point, writeable), (timeout, writeable)}), | |||
| (7, RenewAccepted, {(contract_id, writeable), (offer_per_update_point, writeable), (accept_per_update_point, writeable), (buffer_transaction, writeable), (buffer_script_pubkey, writeable), (timeout, writeable), (own_payout, writeable)}), | |||
| (8, RenewConfirmed, {(contract_id, writeable), (offer_per_update_point, writeable), (accept_per_update_point, writeable), (buffer_transaction, writeable), (buffer_script_pubkey, writeable), (offer_buffer_adaptor_signature, {cb_writeable, write_ecdsa_adaptor_signature, read_ecdsa_adaptor_signature}), (timeout, writeable), (own_payout, writeable), (total_collateral, writeable)}), | |||
| (10, RenewFinalized, {(contract_id, writeable), (offer_per_update_point, writeable), (accept_per_update_point, writeable), (buffer_transaction, writeable), (buffer_script_pubkey, writeable), (offer_buffer_adaptor_signature, {cb_writeable, write_ecdsa_adaptor_signature, read_ecdsa_adaptor_signature}), (accept_buffer_adaptor_signature, {cb_writeable, write_ecdsa_adaptor_signature, read_ecdsa_adaptor_signature}), (timeout, writeable), (own_payout, writeable), (total_collateral, writeable)}), | |||
| (10, RenewFinalized, {(contract_id, writeable), (prev_offer_per_update_point, writeable), (buffer_transaction, writeable), (buffer_script_pubkey, writeable), (offer_buffer_adaptor_signature, {cb_writeable, write_ecdsa_adaptor_signature, read_ecdsa_adaptor_signature}), (accept_buffer_adaptor_signature, {cb_writeable, write_ecdsa_adaptor_signature, read_ecdsa_adaptor_signature}), (timeout, writeable), (own_payout, writeable), (total_collateral, writeable)}), | |||
There was a problem hiding this comment.
Reminder for consumers that this will break any channel stuck in this state prior to this patch. Cc @holzeis.
There was a problem hiding this comment.
you mean any channel that is already stuck?
There was a problem hiding this comment.
Yeah, if a channel is currently in RenewFinalized we won't be able to load it after we update to this patch.
| /// The previous per update point that was used by the offer party for the previous | ||
| /// state of the channel. | ||
| prev_offer_per_update_point: PublicKey, |
There was a problem hiding this comment.
How come we can forget about offer_per_update_point and accept_per_update_point here? And why do we suddenly care about the previous offer per update point?
There was a problem hiding this comment.
IIRC the offer_per_updated_point was not really used and/or redundant with the own_/counter_per_update_point of the SignedChannel struct. The reason for caring about the previous one is to be able to verify that the revocation secret we receive later matches with the point we had.
Tibo-lg
force-pushed
the
fix/ln-dlc/renew-after-disconnect
branch
from
a8070aa to
456e494
Compare
Fix for the bug mentioned here where settling a dlc channel after disconnecting during the opening errors.