(We express it that way rather than 300 zats/1000 bytes, because the
threshold is always rounded to an integer and then multiplied by 3.)
Bitcoin Core added the concept of "dust" in bitcoin/bitcoin#2577.
At that point the dust threshold was tied to three times the
minRelayTxFee rate, with the motivation that if you'd pay more than
a third of the minimum relay fee to spend something, it should be
considered dust. This was implemented as a standard rule rejecting
dust outputs.
This motivation will not apply after ZIP 317 block construction
is implemented: at that point the ZIP 317 marginal fee will be
5000 zats per logical action, but the dust threshold rate will
still be three times 100 zats per 1000 bytes. Those costs would
only coincide if the marginal size per logical action were
5000/300 * 1000 ~= 16667 bytes, and in practice the marginal size
for any kind of input is much smaller than that.
However, to avoid interoperability problems (older wallets creating
transactions that newer nodes will reject because they view the
outputs as dust), we will have to coordinate any increase in the
dust threshold carefully.
More history: in Zcash the minRelayTxFee rate was 5000 zats/1000 bytes
at launch, changed to 1000 zats/1000 bytes in zcashd v1.0.3 and to
100 zats/1000 bytes in zcashd v1.0.7-1 (#2141). The relaying problem
for shielded transactions (#1969) that prompted the latter change was
fixed more thoroughly by the addition of `CFeeRate::GetFeeForRelay`
in #4916, ensuring that a transaction paying `DEFAULT_FEE` can always
be relayed. At the same time the default fee was set to 1000 zats,
per ZIP 313.
An earlier commit in this PR changed relaying policy to be more strict
about enforcing minRelayTxFee. The commit just before this one also
allowed `-minrelaytxfee=0`, which we are going to use to avoid some test
breakage. But if the dust threshold rate were still set to three times
the minRelayTxFee rate, then setting `-minrelaytxfee=0` would have the
side effect of setting the dust threshold to zero, which is not intended.
Bitcoin Core took a different approach to disentangling the dust
threshold from the relay threshold, adding a `-dustrelayfee` option
(bitcoin/bitcoin#9380). We don't want to do that because it is likely
that we will change the dust policy again, and adding a user-visible
config option might conflict with that. Also, it isn't a good idea for
the dust threshold rate to be configurable per node; it's a standard
rule parameter and should only be changed with network-wide coordination
(if it is increased then wallets have to change before nodes, and vice
versa if it is decreased). So for now we set it to a constant that
matches the behaviour before this PR.
Since we can no longer modify the dust threshold, we remove a check
from transaction_tests.cpp that relied on doing so.
This change also indirectly fixes a false-positive assertion error that
would occur in `SpendableInputs::LimitToAmount` if we allowed the dust
threshold to be zero.
Signed-off-by: Daira Emma Hopwood <daira@jacaranda.org>
Remove GetPriority and ComputePriority. Remove internal machinery for tracking priority in CTxMemPoolEntry.
(cherry picked from commit bitcoin/bitcoin@359e8a03d1)
Zcash:
* We don't have `src/bench/mempool_eviction.cpp`.
* We don't have `-walletrejectlongchains`.
* Now we can remove `MAX_PRIORITY`.
* Fix a comment in `coins.h` while we're changing code next to it.
* Update the `Mempool/PriorityStatsDoNotCrash` regression test.
Signed-off-by: Daira Emma Hopwood <daira@jacaranda.org>
zcash/zcash#5987 added a bridge to `orchard::Bundle<Authorized, Amount>`
for `getrawtransaction`. This commit expands it to also cover the
consensus rules, by migrating over missing functionality from the
hand-written FFI methods, and exposing the Orchard `BatchValidator` type
directly (as with Sapling) instead of via the C++ `AuthValidator`
intermediary.
Part of zcash/zcash#6397.
This switches the Merkle tree logic for blocks to one that runs in constant (small) space.
The old code is moved to tests, and a new test is added that for various combinations of
block sizes, transaction positions to compute a branch for, and mutations:
* Verifies that the old code and new code agree for the Merkle root.
* Verifies that the old code and new code agree for the Merkle branch.
* Verifies that the computed Merkle branch is valid.
* Verifies that mutations don't change the Merkle root.
* Verifies that mutations are correctly detected.
(cherry picked from commit bitcoin/bitcoin@eece63fa72)
Assume that when a wallet transaction has a valid block hash and transaction position
in it, the transaction is actually there. We're already trusting wallet data in a
much more fundamental way anyway.
To prevent backward compatibility issues, a new record is used for storing the
block locator in the wallet. Old wallets will see a wallet file synchronized up
to the genesis block, and rescan automatically.
(cherry picked from commit bitcoin/bitcoin@391dff16fe)
Since the wallet ecosystem may not be fully updated to handle
v5 transaction parsing at the point of NU5 activation, some
nodes may prefer to construct V4 transactions when not including
Orchard transaction components.
This change adds a CLI flag that allows node users to specify
that preference.
Remove IncrementalSinsemillaTree; this will be replaced by
a more full-featured OrchardWallet type which embeds the
incremental merkle tree used in wallet operations.
Without this change, mainnet and testnet nodes following standard rules
will not accept v5 transactions into their mempools, causing them to be
ignored.
If `valueBalanceSapling` were `INT64_MIN`, the negation would be undefined
behaviour. Also, if `valueBalanceSapling` were a large-magnitude value
then `nValueOut += -valueBalanceSapling;` could overflow, which would be
undefined behaviour.
In practice neither of these cases can happen because `GetValueOut()` is
only called on non-contextually valid transactions, for which
`valueBalanceSapling` will have been checked to be in range.
`GetShieldedValueIn()` is similarly cleaned up for consistency.
Co-authored-by: Daira Hopwood <daira@jacaranda.org>
We are already checking the other bundle-specific consensus rules:
- Encodings are checked during transaction parsing.
- Signatures are batch-validated.
Closeszcash/zcash#5195.