Fix inbound on-chain payment txid not updating after RBF replacement

[febyeji]

Found via CI failure in the onchain_fee_bump_rbf test:

• When a sender performs multiple RBF fee bumps, the receiver may not update its inbound payment record to reflect the final confirmed replacement transaction. This causes onchain_fee_bump_rbf to fail intermittently:

assertion `left == right` failed: node_a inbound payment txid should be updated to the second replacement txid

Fix

• find_payment_by_txid now also checks the payment_store (not just pending_payment_store) for onchain payments matching the target txid.
• New find_payment_by_conflicting_tx method: if a transaction shares inputs with an existing payment's transaction, it must be an RBF replacement. This works regardless of whether intermediate replacements were observed.

[ldk-reviews-bot]

I've assigned @tnull as a reviewer!
I'll wait for their review and will help manage the review process.
Once they submit their review, I'll check if a second reviewer would be helpful.

[tnull]

cc @Camillarhi

[Camillarhi]

Thanks for this. The fix looks good for the missed-intermediate-bump case, but I think both find_payment_by_txid and find_payment_by_conflicting_tx are reaching into payment_store when pending_payment_store should be enough. Anything still replaceable via RBF hasn't hit ANTI_REORG_DELAY yet, so it's still in pending_payment_store. Payments that have left it are already confirmed and can't be replaced anyway.

[febyeji]

Thanks for the review! I've addressed all three points in dbef121.

[Camillarhi]

Before we proceed, could you point us to the specific CI run where the onchain_fee_bump_rbf failure occurred? I reviewed the recent open PRs and couldn't find a failure matching the assertion you described. A link to the failed run would help us confirm the reproduction and validate the fix.

[ldk-reviews-bot]

🔔 1st Reminder

Hey @tnull! This PR has been waiting for your review.
Please take a look when you have a chance. If you're unable to review, please let us know so we can find another reviewer.

[febyeji]

@Camillarhi The onchain_fee_bump_rbf failure was occurring on the CBF branch CI: https://github.com/lightningdevkit/ldk-node/actions/runs/23131104958/job/67184465048. After investigation, the root cause turned out to be in the CBF branch's ChainTipChanged handler, not in the RBF tracking logic itself. The fix has been applied directly in the CBF branch.

I think this PR can be closed. Thank you very much for the review feedback.