delinvoice was orginally documented to only allow deletion of unpaid
invoices, but there might be reasons to delete paid ones or unexpired ones.
But we have to avoid the race where someone pays as it's deleted: the
easiest way is to have the caller tell us the status, and fail if
it's wrong.
Fixes: #477
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This matches the other names, and also the return value is about to change.
This will be removed before release!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Paid invoices need to know how much was actually paid: both for the case
where no 'msatoshi' amount was specified, and for the normal case, where
clients are permitted to overpay in order to help them disguise their
payments.
While we migrate the db, we leave this field as 0 for old paid
invoices. This is unhelpful for accounting, but at least clearly
indicates what happened if we find this in the wild.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
'rhash' is the old terminology, but 'payment_preimage' and
'payment_hash' were decided on for the BOLTs, so we should fix that here.
We still use rhash internally, but that's much easier to fix.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
1) Note incorrect label in txs in Fig1 of LN paper.
2) "an atomic-swap an on-chain.." -> "an atomic-swap *to* an on-chain"
3) "By using a dual anchor and escape transactions" -> "By using a rebalanced single anchor"
4) References to appendices fixed.
5) Move escape transaction scripts out to escape appendix.
6) Fix URL in bibliography (missing comma).
Reported-by: John Newbery
Closes: #11Closes: #12
Reported-by:
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We can save an opcode in both HTLC send and receive scripts by using
OP_2DROP instead of OP_DROP twice.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
StephenM347 on #lightning-dev indicated it could be clearer that
this is all about changing commitment transactions.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's slightly clearer to do the "2" outside the branch.
Also, spending it requires the extra 0 due to OP_CHECKMULTISIG bug.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>