Olaoluwa Osuntokun
Olaoluwa Osuntokun
> I think we need to do this instead. From caller's view the behavior of the new btcwallet.PublishTransaction is not changed. Thought about this a bit more, and it's potentially...
Here's a crude diff I used to verify the question above: ```diff diff --git a/lnwallet/btcwallet/btcwallet.go b/lnwallet/btcwallet/btcwallet.go index ff0161d36..455af3126 100644 --- a/lnwallet/btcwallet/btcwallet.go +++ b/lnwallet/btcwallet/btcwallet.go @@ -14,7 +14,6 @@ import ( "github.com/btcsuite/btcd/btcutil/hdkeychain"...
Pasting some initial investigation by @carlaKC: I added a timeout to `TestChannelLinkCancelFullCommitment` (see [branch](https://github.com/carlaKC/lnd/tree/TestChannelLinkCancelFullCommitment-deadlock)) and took a mutex profile. I suspect that this may be a deadlock similair to [what...
Related PR: https://github.com/lightningnetwork/lnd/pull/6633
@carlaKC are you still planning on tackling this?
PR up here: #6810. Thx @guggero for starting the impl
So safe to close this issue?
The refactoring I did in #6770 should make it easier to write unit tests for this change as well.
Ok, I think this is g2g now: just pushed up fixes, a new unit test, and also an itest to make sure negotiation works e2e.