lbcwallet/createtx.go

398 lines
12 KiB
Go
Raw Normal View History

/*
* Copyright (c) 2013, 2014 Conformal Systems LLC <info@conformal.com>
*
* Permission to use, copy, modify, and distribute this software for any
* purpose with or without fee is hereby granted, provided that the above
* copyright notice and this permission notice appear in all copies.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
package main
import (
"bytes"
"errors"
"fmt"
badrand "math/rand"
"sort"
"time"
"github.com/conformal/btcchain"
"github.com/conformal/btcscript"
"github.com/conformal/btcutil"
"github.com/conformal/btcwallet/keystore"
"github.com/conformal/btcwallet/txstore"
"github.com/conformal/btcwire"
)
// InsufficientFunds represents an error where there are not enough
// funds from unspent tx outputs for a wallet to create a transaction.
// This may be caused by not enough inputs for all of the desired total
// transaction output amount, or due to
type InsufficientFunds struct {
in, out, fee btcutil.Amount
}
// Error satisifies the builtin error interface.
func (e InsufficientFunds) Error() string {
total := e.out + e.fee
if e.fee == 0 {
return fmt.Sprintf("insufficient funds: transaction requires "+
"%s input but only %v spendable", total, e.in)
}
return fmt.Sprintf("insufficient funds: transaction requires %s input "+
"(%v output + %v fee) but only %v spendable", total, e.out,
e.fee, e.in)
}
// ErrNonPositiveAmount represents an error where a bitcoin amount is
// not positive (either negative, or zero).
var ErrNonPositiveAmount = errors.New("amount is not positive")
// ErrNegativeFee represents an error where a fee is erroneously
// negative.
var ErrNegativeFee = errors.New("fee is negative")
// defaultFeeIncrement is the default minimum transation fee (0.0001 BTC,
// measured in satoshis) added to transactions requiring a fee.
const defaultFeeIncrement = 10000
2013-10-07 21:14:39 +02:00
type CreatedTx struct {
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
tx *btcutil.Tx
changeAddr btcutil.Address
changeIndex int // negative if no change
}
// ByAmount defines the methods needed to satisify sort.Interface to
// sort a slice of Utxos by their amount.
type ByAmount []txstore.Credit
func (u ByAmount) Len() int { return len(u) }
func (u ByAmount) Less(i, j int) bool { return u[i].Amount() < u[j].Amount() }
func (u ByAmount) Swap(i, j int) { u[i], u[j] = u[j], u[i] }
// selectInputs selects the minimum number possible of unspent
// outputs to use to create a new transaction that spends amt satoshis.
// btcout is the total number of satoshis which would be spent by the
// combination of all selected previous outputs. err will equal
// ErrInsufficientFunds if there are not enough unspent outputs to spend amt
// amt.
func selectInputs(eligible []txstore.Credit, amt, fee btcutil.Amount,
minconf int) (selected []txstore.Credit, out btcutil.Amount, err error) {
// Iterate throguh eligible transactions, appending to outputs and
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
// increasing out. This is finished when out is greater than the
// requested amt to spend.
selected = make([]txstore.Credit, 0, len(eligible))
for _, e := range eligible {
selected = append(selected, e)
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
out += e.Amount()
if out >= amt+fee {
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
return selected, out, nil
}
}
if out < amt+fee {
return nil, 0, InsufficientFunds{out, amt, fee}
}
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
return selected, out, nil
}
// txToPairs creates a raw transaction sending the amounts for each
// address/amount pair and fee to each address and the miner. minconf
// specifies the minimum number of confirmations required before an
// unspent output is eligible for spending. Leftover input funds not sent
// to addr or as a fee for the miner are sent to a newly generated
// address. If change is needed to return funds back to an owned
// address, changeUtxo will point to a unconfirmed (height = -1, zeroed
// block hash) Utxo. ErrInsufficientFunds is returned if there are not
// enough eligible unspent outputs to create the transaction.
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
func (w *Wallet) txToPairs(pairs map[string]btcutil.Amount,
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
minconf int) (*CreatedTx, error) {
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
// Key store must be unlocked to compose transaction. Grab the
// unlock if possible (to prevent future unlocks), or return the
// error if the keystore is already locked.
heldUnlock, err := w.HoldUnlock()
if err != nil {
return nil, err
}
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
defer heldUnlock.Release()
// Create a new transaction which will include all input scripts.
msgtx := btcwire.NewMsgTx()
// Calculate minimum amount needed for inputs.
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
var amt btcutil.Amount
for _, v := range pairs {
// Error out if any amount is negative.
if v <= 0 {
return nil, ErrNonPositiveAmount
}
amt += v
}
// Add outputs to new tx.
for addrStr, amt := range pairs {
addr, err := btcutil.DecodeAddress(addrStr, activeNet.Params)
if err != nil {
return nil, fmt.Errorf("cannot decode address: %s", err)
}
// Add output to spend amt to addr.
pkScript, err := btcscript.PayToAddrScript(addr)
if err != nil {
return nil, fmt.Errorf("cannot create txout script: %s", err)
}
txout := btcwire.NewTxOut(int64(amt), pkScript)
msgtx.AddTxOut(txout)
}
// Get current block's height and hash.
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
bs, err := w.chainSvr.BlockStamp()
if err != nil {
return nil, err
}
// Make a copy of msgtx before any inputs are added. This will be
// used as a starting point when trying a fee and starting over with
// a higher fee if not enough was originally chosen.
txNoInputs := msgtx.Copy()
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
unspent, err := w.TxStore.UnspentOutputs()
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
if err != nil {
return nil, err
}
// Filter out unspendable outputs, that is, remove those that (at this
// time) are not P2PKH outputs. Other inputs must be manually included
// in transactions and sent (for example, using createrawtransaction,
// signrawtransaction, and sendrawtransaction).
eligible := make([]txstore.Credit, 0, len(unspent))
for i := range unspent {
switch btcscript.GetScriptClass(unspent[i].TxOut().PkScript) {
case btcscript.PubKeyHashTy:
if !unspent[i].Confirmed(minconf, bs.Height) {
continue
}
// Coinbase transactions must have have reached maturity
// before their outputs may be spent.
if unspent[i].IsCoinbase() {
target := btcchain.CoinbaseMaturity
if !unspent[i].Confirmed(target, bs.Height) {
continue
}
}
// Locked unspent outputs are skipped.
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
if w.LockedOutpoint(*unspent[i].OutPoint()) {
continue
}
eligible = append(eligible, unspent[i])
}
}
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
// Sort eligible inputs, as selectInputs expects these to be sorted
// by amount in reverse order.
sort.Sort(sort.Reverse(ByAmount(eligible)))
var selectedInputs []txstore.Credit
// changeAddr is nil/zeroed until a change address is needed, and reused
// again in case a change utxo has already been chosen.
2014-03-17 16:24:23 +01:00
var changeAddr btcutil.Address
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
var changeIdx int
// Get the number of satoshis to increment fee by when searching for
// the minimum tx fee needed.
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
fee := btcutil.Amount(0)
for {
msgtx = txNoInputs.Copy()
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
changeIdx = -1
// Select eligible outputs to be used in transaction based on the amount
// neededing to sent, and the current fee estimation.
inputs, btcin, err := selectInputs(eligible, amt, fee, minconf)
if err != nil {
return nil, err
}
// Check if there are leftover unspent outputs, and return coins back to
// a new address we own.
change := btcin - amt - fee
if change > 0 {
// Get a new change address if one has not already been found.
if changeAddr == nil {
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
changeAddr, err = w.KeyStore.ChangeAddress(bs)
if err != nil {
return nil, fmt.Errorf("failed to get next address: %s", err)
}
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
w.KeyStore.MarkDirty()
err = w.chainSvr.NotifyReceived([]btcutil.Address{changeAddr})
if err != nil {
return nil, fmt.Errorf("cannot request updates for "+
"change address: %v", err)
}
}
// Spend change.
pkScript, err := btcscript.PayToAddrScript(changeAddr)
if err != nil {
return nil, fmt.Errorf("cannot create txout script: %s", err)
}
msgtx.AddTxOut(btcwire.NewTxOut(int64(change), pkScript))
// Randomize index of the change output.
rng := badrand.New(badrand.NewSource(time.Now().UnixNano()))
r := rng.Int31n(int32(len(msgtx.TxOut))) // random index
c := len(msgtx.TxOut) - 1 // change index
msgtx.TxOut[r], msgtx.TxOut[c] = msgtx.TxOut[c], msgtx.TxOut[r]
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
changeIdx = int(r)
}
// Selected unspent outputs become new transaction's inputs.
for _, ip := range inputs {
msgtx.AddTxIn(btcwire.NewTxIn(ip.OutPoint(), nil))
}
for i, input := range inputs {
// Errors don't matter here, as we only consider the
// case where len(addrs) == 1.
_, addrs, _, _ := input.Addresses(activeNet.Params)
if len(addrs) != 1 {
continue
}
apkh, ok := addrs[0].(*btcutil.AddressPubKeyHash)
if !ok {
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
continue // don't handle inputs to this yet
}
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
ai, err := w.KeyStore.Address(apkh)
if err != nil {
return nil, fmt.Errorf("cannot get address info: %v", err)
}
pka := ai.(keystore.PubKeyAddress)
privkey, err := pka.PrivKey()
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
if err != nil {
return nil, fmt.Errorf("cannot get private key: %v", err)
}
sigscript, err := btcscript.SignatureScript(msgtx, i,
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
input.TxOut().PkScript, btcscript.SigHashAll, privkey,
ai.Compressed())
if err != nil {
return nil, fmt.Errorf("cannot create sigscript: %s", err)
}
msgtx.TxIn[i].SignatureScript = sigscript
}
noFeeAllowed := false
if !cfg.DisallowFree {
noFeeAllowed = allowFree(bs.Height, inputs, msgtx.SerializeSize())
}
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
if minFee := minimumFee(w.FeeIncrement, msgtx, noFeeAllowed); fee < minFee {
fee = minFee
} else {
selectedInputs = inputs
break
}
}
// Validate msgtx before returning the raw transaction.
2013-10-25 21:21:38 +02:00
flags := btcscript.ScriptCanonicalSignatures
2013-10-11 16:53:55 +02:00
bip16 := time.Now().After(btcscript.Bip16Activation)
2013-10-25 21:21:38 +02:00
if bip16 {
flags |= btcscript.ScriptBip16
}
for i, txin := range msgtx.TxIn {
engine, err := btcscript.NewScript(txin.SignatureScript,
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
selectedInputs[i].TxOut().PkScript, i, msgtx, flags)
if err != nil {
return nil, fmt.Errorf("cannot create script engine: %s", err)
}
if err = engine.Execute(); err != nil {
return nil, fmt.Errorf("cannot validate transaction: %s", err)
}
}
buf := bytes.Buffer{}
buf.Grow(msgtx.SerializeSize())
if err := msgtx.BtcEncode(&buf, btcwire.ProtocolVersion); err != nil {
// Hitting OOM by growing or writing to a bytes.Buffer already
// panics, and all returned errors are unexpected.
panic(err)
}
info := &CreatedTx{
Remove account support, fix races on btcd connect. This commit is the result of several big changes being made to the wallet. In particular, the "handshake" (initial sync to the chain server) was quite racy and required proper synchronization. To make fixing this race easier, several other changes were made to the internal wallet data structures and much of the RPC server ended up being rewritten. First, all account support has been removed. The previous Account struct has been replaced with a Wallet structure, which includes a keystore for saving keys, and a txstore for storing relevant transactions. This decision has been made since it is the opinion of myself and other developers that bitcoind accounts are fundamentally broken (as accounts implemented by bitcoind support both arbitrary address groupings as well as moving balances between accounts -- these are fundamentally incompatible features), and since a BIP0032 keystore is soon planned to be implemented (at which point, "accounts" can return as HD extended keys). With the keystore handling the grouping of related keys, there is no reason have many different Account structs, and the AccountManager has been removed as well. All RPC handlers that take an account option will only work with "" (the default account) or "*" if the RPC allows specifying all accounts. Second, much of the RPC server has been cleaned up. The global variables for the RPC server and chain server client have been moved to part of the rpcServer struct, and the handlers for each RPC method that are looked up change depending on which components have been set. Passthrough requests are also no longer handled specially, but when the chain server is set, a handler to perform the passthrough will be returned if the method is not otherwise a wallet RPC. The notification system for websocket clients has also been rewritten so wallet components can send notifications through channels, rather than requiring direct access to the RPC server itself, or worse still, sending directly to a websocket client's send channel. In the future, this will enable proper registration of notifications, rather than unsolicited broadcasts to every connected websocket client (see issue #84). Finally, and the main reason why much of this cleanup was necessary, the races during intial sync with the chain server have been fixed. Previously, when the 'Handshake' was run, a rescan would occur which would perform modifications to Account data structures as notifications were received. Synchronization was provided with a single binary semaphore which serialized all access to wallet and account data. However, the Handshake itself was not able to run with this lock (or else notifications would block), and many data races would occur as both notifications were being handled. If GOMAXPROCS was ever increased beyond 1, btcwallet would always immediately crash due to invalid addresses caused by the data races on startup. To fix this, the single lock for all wallet access has been replaced with mutexes for both the keystore and txstore. Handling of btcd notifications and client requests may now occur simultaneously. GOMAXPROCS has also been set to the number of logical CPUs at the beginning of main, since with the data races fixed, there's no reason to prevent the extra parallelism gained by increasing it. Closes #78. Closes #101. Closes #110.
2014-07-09 05:17:38 +02:00
tx: btcutil.NewTx(msgtx),
changeAddr: changeAddr,
changeIndex: changeIdx,
}
return info, nil
}
// minimumFee calculates the minimum fee required for a transaction.
// If allowFree is true, a fee may be zero so long as the entire
// transaction has a serialized length less than 1 kilobyte
// and none of the outputs contain a value less than 1 bitcent.
// Otherwise, the fee will be calculated using TxFeeIncrement,
// incrementing the fee for each kilobyte of transaction.
func minimumFee(incr btcutil.Amount, tx *btcwire.MsgTx, allowFree bool) btcutil.Amount {
txLen := tx.SerializeSize()
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
fee := btcutil.Amount(int64(1+txLen/1000) * int64(incr))
if allowFree && txLen < 1000 {
fee = 0
}
if fee < incr {
for _, txOut := range tx.TxOut {
if txOut.Value < btcutil.SatoshiPerBitcent {
return incr
}
}
}
2014-07-08 18:22:09 +02:00
if fee < 0 || fee > btcutil.MaxSatoshi {
fee = btcutil.MaxSatoshi
}
return fee
}
// allowFree calculates the transaction priority and checks that the
2014-07-08 18:22:09 +02:00
// priority reaches a certain threshold. If the threshhold is
// reached, a free transaction fee is allowed.
func allowFree(curHeight int32, txouts []txstore.Credit, txSize int) bool {
2014-07-08 18:22:09 +02:00
const blocksPerDayEstimate = 144.0
const txSizeEstimate = 250.0
const threshold = btcutil.SatoshiPerBitcoin * blocksPerDayEstimate / txSizeEstimate
var weightedSum int64
for _, txout := range txouts {
Another day, another tx store implementation. The last transaction store was a great example of how not to write scalable software. For a variety of reasons, it was very slow at processing transaction inserts. Among them: 1) Every single transaction record being saved in a linked list (container/list), and inserting into this list would be an O(n) operation so that records could be ordered by receive date. 2) Every single transaction in the above mentioned list was iterated over in order to find double spends which must be removed. It is silly to do this check for mined transactions, which already have been checked for this by btcd. Worse yet, if double spends were found, the list would be iterated a second (or third, or fourth) time for each removed transaction. 3) All spend tracking for signed-by-wallet transactions was found on each transaction insert, even if the now spent previous transaction outputs were known by the caller. This list could keep going on, but you get the idea. It was bad. To resolve these issues a new transaction store had to be implemented. The new implementation: 1) Tracks mined and unmined transactions in different data structures. Mined transactions are cheap to track because the required double spend checks have already been performed by the chain server, and double spend checks are only required to be performed on newly-inserted mined transactions which may conflict with previous unmined transactions. 2) Saves mined transactions grouped by block first, and then by their transaction index. Lookup keys for mined transactions are simply the block height (in the best chain, that's all we save) and index of the transaction in the block. This makes looking up any arbitrary transaction almost an O(1) operation (almost, because block height and block indexes are mapped to their slice indexes with a Go map). 3) Saves records in each transaction for whether the outputs are wallet credits (spendable by wallet) and for whether inputs debit from previous credits. Both structures point back to the source or spender (credits point to the transaction that spends them, or nil for unspent credits, and debits include keys to lookup the transaction credits they spent. While complicated to keep track of, this greatly simplifies the spent tracking for transactions across rollbacks and transaction removals. 4) Implements double spend checking as an almost O(1) operation. A Go map is used to map each previous outpoint for all unconfirmed transactions to the unconfirmed tx record itself. Checking for double spends on confirmed transaction inserts only involves looking up each previous outpoint of the inserted tx in this map. If a double spend is found, removal is simplified by only removing the transaction and its spend chain from store maps, rather than iterating a linked list several times over to remove each dead transaction in the spend chain. 5) Allows the caller to specify the previous credits which are spent by a debiting transaction. When a transaction is created by wallet, the previous outputs are already known, and by passing their record types to the AddDebits method, lookups for each previously unspent credit are omitted. 6) Bookkeeps all blocks with transactions with unspent credits, and bookkeeps the transaction indexes of all transactions with unspent outputs for a single block. For the case where the caller adding a debit record does not know what credits a transaction debits from, these bookkeeping structures allow the store to only consider known unspent transactions, rather than searching through both spent and unspents. 7) Saves amount deltas for the entire balance as a result of each block, due to transactions within that block. This improves the performance of calculating the full balance by not needing to iterate over every transaction, and then every credit, to determine if a credit is spent or unspent. When transactions are moved from unconfirmed to a block structure, the amount deltas are incremented by the amount of all transaction credits (both spent and unspent) and debited by the total amount the transaction spends from previous wallet credits. For the common case of calculating a balance with just one confirmation, the only involves iterating over each block structure and adding the (possibly negative) amount delta. Coinbase rewards are saved similarly, but with a different amount variable so they can be seperatly included or excluded. Due to all of the changes in how the store internally works, the serialization format has changed. To simplify the serialization logic, support for reading the last store file version has been removed. Past this change, a rescan (run automatically) will be required to rebuild the transaction history.
2014-05-05 23:12:05 +02:00
depth := chainDepth(txout.BlockHeight, curHeight)
weightedSum += int64(txout.Amount()) * int64(depth)
}
priority := float64(weightedSum) / float64(txSize)
2014-07-08 18:22:09 +02:00
return priority > threshold
}
// chainDepth returns the chaindepth of a target given the current
// blockchain height.
func chainDepth(target, current int32) int32 {
if target == -1 {
// target is not yet in a block.
return 0
}
// target is in a block.
return current - target + 1
}