The comment says "only allow recent nodes (10mins) after we failed 30 times",
but the server actually did the opposite and allowed only recent nodes before
30 failed connection attempts. This corrects the server's behavior.
This commit modifies the createTxRawResult code path along with callers
to work with block headers as opposed to btcutil.Blocks. This in turn
allows the code in handleGetRawTransaction and
handleSearchRawTransactions to perform a much cheaper block header load
as opposed to a full block load.
While here, also very slightly optimize the createVinList function to
avoid creating a util.Tx wrapper and to take advantage of the
btcutil.Amount type added after the function was originally written
This commit updates the merkle block handling to for the latest changes
to the btcutil API and optimizes it along the way.
Previously, the code was inefficiently reloading the transactions for
the matched hashes from the database instead of simply pulling them from
the full block that was used to create the merkle block.
At the current time, there is no difference between the wire encoding
at protocol version 0 and the stable long-term storage format. These
methods are simply for consistency with the other types.
This commit updates the wire tests for transactions which force
serialization and deserialization errors to force an error in the the
transaction lock time path.
This brings the wire test coverage back up to 100%.
This commit correctly replaces persistent peers that are being retried in
the list of persistent peers so it will continue to be retried as
intended.
Also, limit the maximum retry interval for persistent peers to 5 minutes.
Fixes#463.
This change moves IsFinalizedTransaction to txscript and also changes
the first argument to take a wire.MsgTx instead of btcutil.Tx. This
is needed for an upcoming diff in which txscript will require
IsFinalizedTransaction and we do not want to import the btcd/blockchain.
This allows clients watching for both to know when all mined
transaction notifications for the block have been received.
Otherwise, clients would be aware that there is a new block, see the
exact same block hash/height in the next tx notifications, but never
know when all txs from the block have been received and processed.
This fixes a synchronization issue in btcwallet where the wallet would
mark itself synced to some block height before any of those blocks'
transactions were processed. If the RPC connection is lost between
sending the blockconnected notification and receiving the last
transaction notification, the wallet would not know of this and
continue with missing transactions.
The configured onion proxy was not being used due to checking if the
passed string had suffix '.onion', which never matched because the
port number is part of the string.
This commit adds an additional step to the README.md install section to
run the go version command and check the version so people that are
installing it for the first time and ensure they are running a high
enough version and have GOROOT and GOPATH set correctly.
This commit corrects an issue where the reject message sent in response
to rejected transactions had the Cmd field set to "block" instead of
"tx".
Fixes#436.
Tor stream isolation randomizes proxy user credentials resulting in
Tor creating a new circuit for each connection. This makes it more
difficult to correlate connections.
Idea from Wladimir J. van der Laan via Bitcoin Core.
The option -maxorphantx allows the user to specify the number of
orphan transactions to keep in memory.
Also, lower the default max orphan count from 10000 to 1000.
This commit refactors the consensus rule checks for block headers and
blocks in the blockchain package into separate functions. These changes
contain no modifications to consensus rules and the code still passes all
block consensus tests. It is only a refactoring.
This is being done to help pave the way toward supporting concurrent
downloads. While the package already supports headers-first mode up
through the latest checkpoint through the use of the BFFastAdd flag and
hard-coded checkpoints, it currently only works when downloading from a
single peer. In order to support concurrent downloads from multiple
peers, the ability for the caller to do things such as independently
checking a block header (both context-free and full-context checks) will
be needed.
There are several more changes that will be necessary to support
concurrent downloads as well, such as making the package concurrent safe,
modifying it to make use of the new database API, etc. Those changes are
planned for future commits.
The following changes were made to ListTransactionsResult (which
models the long result format used by listtransactions,
listsinceblock, etc.):
- Fee made optional (float64 -> *float64 + omitempty)
- BlockIndex made optional (int64 + omitempty -> *int64 + omitempty)
- InvolvesWatchOnly added (bool + omitempty)
- Vout added (uint32)
The following changes were made to GetTransactionDetailsResult (which
models the short result format of listtransactions):
- InvolvesWatchOnly added (bool + omitempty)
- Fee added (*float64 + omitempty)
- Vout added (uint32)
The combination of pointer types and the omitempty struct tag allow
excluding the field from the JSON object, or including it with the
zero value. This is useful in particular for the fee fields, which
should be included whenever the category is "send" even if the fee is
zero. Other optional fields which are only added to the result object
with non-zero values (such as includeswatchonly) can be reduced to
simply an omitempty tag without the pointer type.
This commit contains fixes from the results of a thorough audit of
txscript to find any cases of script evaluation which doesn't match the
required consensus behavior. These conditions are fairly obscure and
highly unlikely to happen in any real scripts, but they could have
nevertheless been used by a clever attacker with malicious intent to
cause a fork.
Test cases which exercise these conditions have been added to the
reference tests and will contributed upstream to improve the quality for
the entire ecosystem.
Unlike OP_IF and OP_NOTIF which interpret the top stack item as a
number, OP_IFDUP interprets it as a boolean. This has important
consequences because numbers are imited to int32s while booleans can be
an arbitrary number of bytes.
The offending script was found and reported by Jonas Nick through the
use of fuzzing.
