Early sample testing showed multiple users hitting
EWOULDBLOCK/EAGAIN waiting for an initial response from Bitcoin
Core while it was doing some long operation (eg UTXO cache
flushing). Instead of only waiting 5 seconds for each attempt, we
now wait a full two minutes, but only for the first header
response, not each byte.
Our enforced requirements for HTLC acceptance is that we have at
least HTLC_FAIL_BACK_BUFFER blocks before the HTLC expires. When we
receive an HTLC, the HTLC would be "already expired" if its
`cltv_expiry` is current-block + 1 (ie the next block could
broadcast the commitment transaction and time out the HTLC). From
there, we want an extra HTLC_FAIL_BACK_BUFFER in blocks, plus an
extra block or two to account for any differences in the view of
the current height before send or while the HTLC is transiting the
network.
This increases the CLTV_CLAIM_BUFFER constant to 18, much better
capturing how long it takes to go on chain to claim payments.
This is also more in line with other clients, and the spec, which
sets the default CLTV delay in invoices to 18.
As a side effect, we have to increase MIN_CLTV_EXPIRY_DELTA as
otherwise as are subject to an attack where someone can hold an
HTLC being forwarded long enough that we *also* close the channel
on which we received the HTLC.
In #797, we stopped enforcing that read/sent node_announcements
had their addresses sorted. While this is fine in practice, we
should still make a best-effort to sort them to comply with the
spec's forward-compatibility requirements, which we do here in the
ChannelManager.
Since InvoiceFeatures are an implementation detail of InvoiceBuilder, an
explicit call is needed to support the basic_mpp feature. Since it is
dependent on the payment_secret feature, conditionally define the
builder's method only when payment_secret has been set.
Instead of relying on users to set an invoice's features correctly,
enforce the semantics inside InvoiceBuilder. For instance, if the user
sets a PaymentSecret then InvoiceBuilder should ensure the appropriate
feature bits are set. Thus, for this example, the TaggedField
abstraction can be retained while still ensuring BOLT 11 semantics at
the builder abstraction.
Current Bitcoin Core's policy will reject a p2wsh as a dust if it's
under 330 satoshis. A typical p2wsh output is 43 bytes big to which
Core's `GetDustThreshold()` sums up a minimal spend of 67 bytes (even
if a p2wsh witnessScript might be smaller). `dustRelayFee` is set
to 3000 sat/kb, thus 110 * 3000 / 1000 = 330. As all time-sensitive
outputs are p2wsh, a value of 330 sat is the lower bound desired
to ensure good propagation of transactions. We give a bit margin to
our counterparty and pick up 660 satoshis as an accepted
`dust_limit_satoshis` upper bound.
As this reasoning is tricky and error-prone we hardcode it instead of
letting the user picking up a non-sense value.
Further, this lower bound of 330 sats is also hardcoded as another constant
(MIN_DUST_LIMIT_SATOSHIS) instead of being dynamically computed on
feerate (derive_holder_dust_limit_satoshis`). Reducing risks of
non-propagating transactions in casee of failing fee festimation.
While this is less readable, I spent way too long trying to adapt
the bindings generation code to handle glob imports and concluded
it would take refactoring almost the entire import-resolution
logic. While this may be a good refactor to do eventually, its
probably not worth it today.
The C bindings generator now looks to default generic types as the
way to map a struct or enum parameter. Because SignOrCreationError
is only used directly with an error type of `()`, we set that to
the default and assume no other error types are needed.
