#4227 Followups #4280

Activity

1. TheBlueMatt

   added this to the 0.3 milestone on Dec 11, 2025
2. TheBlueMatt commented on Jan 14, 2026

   

   TheBlueMatt

   on Jan 14, 2026

   CollaboratorAuthor

   More actions

   Oh, one more issue I think - the dedup code (also after #4289) doesn't consider things in the holding cells. ChannelMonitors are really only authoritative on the HTLC state if the channel is closed, not before. So because we're looking at the monitors we may double-forward still, I believe. Instead, we really need to dedup on the Channel itself. Now, maybe when we are persisting the Channel in the monitors (and are not writing the holding cell, which I think makes sense) just looking at the monitors will be right, but today it is not.
3. valentinewallace

   self-assigned this

   on Jan 15, 2026
4. TheBlueMatt

   closed this as completedin #4303on Feb 10, 2026