Resolved hashes in payments do not reflect back their settled state

[alexbosworth]

Background

When sending a payment that did not resolve immediately, the eventual preimage revelation via an on-chain HTLC success transaction did not reflect as a success in listpayments.

Your environment

• 0.5.0-beta commit=b341dea3733cb721d6aecefa54123b5b1ac7b126
• MacOS Mojave
• Bitcoin Core 0.17.0
• Sending to lnd 0.5.0

Steps to reproduce

1. Sent to a destination that held on to the HTLC and did not resolve it
2. Restarted my lnd
3. Waited for eventual resolution of the HTLC which happened on chain
4. Settlement message appeared in logs, but no record was added to listpayments of the payment

Expected behaviour

Payments should be listed in listpayments when funds have been successfully transfered

Actual behaviour

A message indicating completion of the HTLC comes in, but there is an error regarding a payment lookup.

2018-10-09 22:45:43.943 [DBG] HSWC: ChannelLink(544281:1936:0): settle-fail-filter &{1 [0]}
2018-10-09 22:45:43.943 [DBG] HSWC: Closed completed SETTLE circuit for b5d5bf498e1a8ba0a3e8d05d5d666e00f2943c9f498dfb6a92af8795d98ceaa2: (0:0:0, 19003) <-> (544281:1936:0, 27)
2018-10-09 22:45:43.955 [DBG] HSWC: Tearing down open circuit with SETTLE pkt, removing circuit=(Chan ID=0:0:0, HTLC ID=19003) with keystone=(Chan ID=544281:1936:0, HTLC ID=27)
2018-10-09 22:45:43.967 [DBG] HSWC: Closed completed SETTLE circuit for b5d5bf498e1a8ba0a3e8d05d5d666e00f2943c9f498dfb6a92af8795d98ceaa2: (0:0:0, 19003) <-> (544281:1936:0, 27)
2018-10-09 22:45:43.967 [ERR] HSWC: Cannot find pending payment with ID 19003

Payment b5d5bf498e1a8ba0a3e8d05d5d666e00f2943c9f498dfb6a92af8795d98ceaa2 is absent from listpayments

[treygriffith]

@halseth is this issue now resolved with #2761 and #2762 ? Or is there something else that still needs to be fixed?

[alexbosworth]

@treygriffith do you still see this issue if you follow the above repro steps? It should be resolved in master

[halseth]

This should indeed be fixed in master. I put up a new test case to be sure, and it is passing: #3194