invoices/invoiceregistry: generate random payaddr for keysend #4343
Conversation
6b2d498 to
bee90cf
Compare
invoices/invoiceregistry_test.go
Outdated
| // Replaying the same context should also succeed to ensure idempotency. | ||
| // This should be handled later in the pipeline by comparing the circuit | ||
| // keys. | ||
| err = ctx.registry.processKeySend(keysendCtx) |
There was a problem hiding this comment.
I think that testing on the external interface of invoice registry is more robust.
There was a problem hiding this comment.
my thought here was that this is a better place to assert the behavior of processKeySend wrt to how it filters errors
There was a problem hiding this comment.
Why is it better? We also already have a parameterized TestKeysend that tests on the external interface. Maybe that can be extended easily.
There was a problem hiding this comment.
afaict you can't test the idempotency of processKeySend because it settles the keysend immediately after
There was a problem hiding this comment.
moved everything to a white box test, but we are discussing holding off on this change and fixing as part of #4338
bee90cf to
46ba6eb
Compare
|
@joostjager rebased on #4334 |
This commit automatically generates a random payment address for any new keysend payments. With the addition of the payment address index, we now enforce uniqueness of payment addresses across all invoices. As a result, if you run master currently you can only accept one keysend payment since they all share a zero-value payment address.
46ba6eb to
7d1c528
Compare
|
Rolled into #4338 |
This commit automatically generates a random payment address for any new
keysend payments. With the addition of the payment address index, we now
enforce uniqueness of payment addresses across all invoices. As a
result, if you run master currently you can only accept one keysend
payment since they all share a zero-value payment address.
Follow up from #4285