If we switch the knuth shuffle to the version that swaps the element
with an element between it and the end of the array, then once we have
gotten to the amount of elements we need they won't change later in the
algorithm. Terminating here means that we only do 23% of the length of
the array worth of random swaps at most.
We make ka.na immutable in the address manager. Whenever we would update
the structure we replace it with a new copy. This beats making a copy of
all addresses once per getaddr command (max is just over 23000 we would
be copying, compared to at most 2000 copies on a new getaddr that has
all addresses we know with newer dates).
On unknown inventory types, handleGetDataMsg would loop forever.
After fixing that, if a getdata request only had unknown inventory
types, it would block forever.
ok @davecgh
Copying the RIPEMD160 after SHA256 hash result into a new stack array
to be used as a map lookup key can be quite expensive, and this should
be avoided if possible on intensive tasks such as rescans. This
change takes advantage of the new Hash160 methods of the
AddressPubKeyHash and AddressScriptHash types to use the address's
underlying hash array directly, rather than creating a copy from the
ScriptAddress result.
Unfortunately, for AddressPubKey, ScriptAddress may return either a
byte slice of len 33 or 65 depending on whether the pubkey is
compressed or not, so no such straightforward optimization is
possible.
As a result of this change, I have seen rescans perform roughly 3.5x
faster than before.
The websocket extension command to register for notifications when a new
transaction has been accepted to the memory pool and the resulting
notifications have been renamed. This commit catches up to the change.
Since a chain verification can take a long time depending on the
parameters, this commit adds a debug print to the RPC server at the info
level for how many blocks are being verified and at what level.
The logic was also slightly modified so the number of blocks being checked
can easily be calculated and shown.
This commit modifies peers to use a max protocol version that is specified
as a constant in the peer code as opposed to the btcwire.ProtocolVersion
constant.
This allows btcwire to be updated to support new protocol versions without
causing peers to claim they support a protocol version which they actually
don't.
This change periodically (about every 10 seconds) notifies the
connected websocket client of the height of the last processed block
as part of the rescan. This enables clients to listen for the
notification and keep track of how much progress a rescan has made
even without any results being found. If the websocket connection is
lost during the rescan, on reconnect, clients may safely start over at
the last notified height.
This commit cleans up and moves a couple of comments in the recent pull
request which implements a rebroadcast handler (#114) in order to avoid
discussing internal state in the exported function comment.
How a function actually accomplishes the stated functionality is not
something that a caller is concerned with. The details about the internal
state are better handled with comments inside the function body.
This commit implements a rebroadcast handler which deals with
rebroadcasting inventory at a random time interval between 0 and 30
minutes. It then uses the new rebroadcast logic to ensure transactions
which were submitted via the sendrawtransaction RPC are rebroadcast until
they make it into a block.
Closes#99.
Rather than using the deprecated TxShas function on a btcutil.Block,
convert handleGetBlock to use the newer preferred method of ranging over
the Transactions to obtain the cached hash of each transaction.
This is a little more efficient since it can avoid creating and caching an
extra slice to keep the hashes in addition to having the hash cached with
each transaction.
This commit adds a new function named NewBlockTemplate along with
supporting infrastructure which is part of the core functionality needed
to support mining.
In particular the function creates a new block template which contains a
fully populated block with a zero nonce that is ready to be solved as well
as additional information regarding the fees and number of signature
operations for each transaction included in the block. The specific
transaction selection logic mirrors the reference implementation.
Various cleanup, optimizations, and comment suggestions provided by
@owainga. Also contains some naming suggestions and comment fixes from
@flammit.
This change adds select statements to each channel write that may
occur from a non-RPCS goroutine to unblock once the server has begun
shutting down. This prevents issues in clean shutdown where non-RPCS
goroutines may block indefinitely on message sends to the RPC server.
This change adds rescan fast paths for all current implementations of
the btcutil.Address interface. Rather than encoding a payment address
string from a txout script, the minimum number of bytes (which depends
on the address type) are copied into a stack array and used as a
lookup key to a map exclusive for that address type. This performs
better than the previous implementation due to avoiding the base58
encoding and generating less garbage (so each GC cycle completes
faster).
If the address type switch falls into the default (unhandled) case,
the encoded payment address is used as a fallback. This keeps the
intended rescan behavior even when new implementations of
btcutil.Address are added without a rescan fast path.
Benchmarks indicate that when a fast path is followed, for 20 byte
RIPEMD-160 hashes (such as P2PKH and P2SH), skipping the payment
address encoding (and thus bypassing the base58 encoding code) can
result in map lookups up to 60x faster.
Rather than returning an error when creating the RPC server an it can't
listen on any of the specified interfaces, only error when it can't listen
on all of the specified interfaces.
Rather than updating the new chain state with the hash and height of the
block that was just processed, query the database for the best block.
This is needed because the block that was just processed might be a side
chain block or have caused a reorg.
This commit introduces a chain state that is updated as blocks are
processed into the block chain instance associated with the block manager.
This has been done because btcchain is currently not safe for concurrent
access and the block manager is typically quite busy processing block and
inventory. This approach allows fast access to most chain information in
a concurrent safe fashion.
Rather than doing a lot of if/else in the configuration load for btcctl,
set the default for the RPC server to localhost and the RPC cert to the
btcd cert, then override them as needed depending on the --wallet and
--testnet flags.
The code was not resetting the counters when logging a message, so the
stats kept increasing rather than showing the number of items processed
since the last log message.
This has been resolved by copying the log function directly from btcd and
modifying it slightly to work with the importer.
blockImporter now has these new properties:
txProcessed - number of transactions processed by the importer
lastHeight - height of the last imported block
lastBlockTime - block time of the last imported block
lastLogTime - last timestamp of when a progress message was shown
the p/progress option in addblocks now indicates how often you want the progress message to appear
