Andreas Schildbach
504b67c186
ByteUtils: check range of values supplied to write helpers for the Java Unsigned Integer API
...
These checks uncovered various uses of illegal values (now fixed).
They should help this won't happen in future.
2023-03-27 17:53:26 +02:00
Sean Gilligan
2aa764d403
VarInt: simplify encode()
2023-03-27 17:38:58 +02:00
Andreas Schildbach
3746516859
VarInt: use constants for possible sizes
2023-03-27 16:54:00 +02:00
Andreas Schildbach
140912b080
VarIntTest: add test for reading and writing
2023-03-27 16:49:21 +02:00
Andreas Schildbach
21fae58bc8
VarInt: implement toString()
2023-03-27 16:48:47 +02:00
Andreas Schildbach
301c8e4a76
VarInt: implement equals()
and hashCode()
...
`originallyEncodedSize` is not considered on purpose, because it is only
relevant for serialization.
2023-03-27 16:26:16 +02:00
Andreas Schildbach
2faaf9febc
VarInt: direct use of ByteBuffer
for reading and writing
2023-03-27 16:00:37 +02:00
Sean Gilligan
b4abf7d794
VarInt: get rid of code duplication in deprecated constructor
2023-03-27 15:50:13 +02:00
Andreas Schildbach
b528fb5180
build.gradle: update SLF4J to 2.0.7
...
Starting with SLF4J 2.0.0, we can use the fluent logger API.
2023-03-27 15:25:02 +02:00
Sean Gilligan
2e3e6d4485
TransactionBroadcast: add awaitSent()
method
...
This just returns the used internally (and therefore tested) `sentFuture`.
2023-03-27 12:28:27 +02:00
Sean Gilligan
bce7b11038
Script: downsize MAX_SCRIPT_ELEMENT_SIZE
to int
from long
...
* The value is 520
* Java collection sizes are `int`
* All usages are as `int`
2023-03-27 09:54:42 +02:00
Sean Gilligan
ee74f34dcc
FullBlockTestGenerator: use writeInt64LE()
for writing a 64-bit integer
...
Previously two `writeInt32LE()` and a right-shift on an `int` by 32 were used.
2023-03-27 09:48:01 +02:00
Sean Gilligan
37a06b150f
Transaction: replace Guava Lists.reversed()
with JDK equivalent
2023-03-27 09:38:38 +02:00
Andreas Schildbach
c5faecf3b6
SegwitAddress: flip arguments of new static constructor fromBech32()
...
Otherwise the deprecation cycle won't be of much use for consumers who pass `null` for `params`.
2023-03-27 09:23:04 +02:00
Andreas Schildbach
150c911750
LegacyAddress: flip arguments of new static constructor fromBase58()
...
Otherwise the deprecation cycle won't be of much use for consumers who pass `null` for `params`.
2023-03-27 09:17:37 +02:00
Andreas Schildbach
4bdb0f85e4
BitcoinURI: flip arguments of new constructor
...
Otherwise the deprecation cycle won't be of much use for consumers who pass `null` for `params`.
2023-03-27 09:14:11 +02:00
Sean Gilligan
884072c9e3
WalletAppKit: make Closeable
2023-03-26 19:40:28 +02:00
Andreas Schildbach
c10cf7ccbe
Sha256Hash: direct use of ByteBuffer
in hashCode()
...
A hash code is just an arbitrary number. It doesn't care about signedness and endianess.
2023-03-26 19:07:22 +02:00
Andreas Schildbach
d478eb2a55
Network: migrate legacyAddressHeader()
and legacyP2SHHeader()
from NetworkParameters
2023-03-26 18:59:51 +02:00
Andreas Schildbach
ba72884832
Transaction: add back a byte[]
-based constructor as deprecated
2023-03-26 18:53:06 +02:00
Sean Gilligan
c05fefe8f8
MarriedKeyChain: simplify Builder.followingKeys()
...
* Add `followingKey()` method that takes a single key
* Deprecate unused `followingKeys(DeterministicKey, DeterministicKey...)`
* Replace usage of Guava's `Lists.asList()`
2023-03-26 18:49:32 +02:00
Sean Gilligan
3e1855b201
SendRequest: make field tx
immutable
2023-03-26 18:03:56 +02:00
Andreas Schildbach
4bdb8d7836
CheckpointManager: fix wrong unit of time
...
This mistake was introduced in commit 838b12c046
.
The original line before the `java.time` migration was:
`60 * 60 * 24 * 7; // one week in seconds`
2023-03-26 17:58:10 +02:00
Sean Gilligan
795da8f54b
Wallet: replace Guava newArrayList()
with singletonList()
2023-03-26 11:16:09 +02:00
Andreas Schildbach
b585f9f08e
KeyChainGroupTest: reduce false positive rate for the bloom filters
...
This will hopefully reduce spurious test failures.
2023-03-25 13:29:20 +01:00
Andreas Schildbach
3bbbf40fa4
Message: add back deprecated method unsafeBitcoinSerialize()
2023-03-25 13:19:59 +01:00
Andreas Schildbach
c581dfb791
ByteUtils: rework JavaDoc of all writeInt*()
helpers
2023-03-24 19:35:52 +01:00
Andreas Schildbach
47a5d73035
ByteUtils: rename all writeUInt*()
helpers to writeInt*()
...
This creates overloads in some cases.
2023-03-24 19:34:06 +01:00
Sean Gilligan
f3c4b5fcf4
BloomFilter: use signed integer for the nonce (nTweak)
...
It's arbitrary data and we're not supposed to do math on it, so it's
safe to use a signed Java `int`.
Also use `new Random().nextInt()` to generate the nonce. We were
previously generating a random double, multiplying by `Long.MAX_VALUE`
and then truncating the low 32-bits of the long. Using `.nextInt()`
might actually generate MORE randomness since there will be no truncation.
2023-03-24 18:34:09 +01:00
Andreas Schildbach
b39ad1904d
Message: remove check against max message size in checkReadLength()
...
Checking against the remaining bytes in the payload should be enough.
The payload can't be longer than max message size.
2023-03-24 16:15:21 +01:00
Andreas Schildbach
9b3b11fd0f
Message: allow parse()
to throw BufferUnderflowException
...
The exception is now handled by the one caller to `parse()`, rather than
by various helpers.
2023-03-24 16:08:17 +01:00
Sean Gilligan
08e8c0c022
BloomFilterTest: use hex representation for constant
...
This avoids the `L` and also shows more clearly the
bit-pattern being used.
2023-03-24 15:44:34 +01:00
Andreas Schildbach
48f4d143f3
FeeFilterMessage: disallow negative fee rate
...
Fee rates are signed "int64_t", according to BIP-133. The spec doesn't
mention, but it goes without saying that negative fee rates make no
sense.
2023-03-24 15:32:54 +01:00
Andreas Schildbach
31d98af8f6
TransactionOutput: disallow negative values
...
Negative values obviously make no sense, except for `-1` which is
used as a sentinel value when calculating `SIGHASH_SINGLE` signatures,
so unfortunately we have to allow that here.
2023-03-24 15:30:02 +01:00
Andreas Schildbach
5bb1e123b6
VersionMessage: serialize negative time
values
...
Surprisingly, version messages serialize time as a `int64_t` (signed 64-bit integer).
2023-03-24 15:26:10 +01:00
Andreas Schildbach
7a0939f7e4
Ping: require a nonce value
...
Pings without nonce are not protocol compliant since version 60001.
Our minimum version is at 70000.
2023-03-24 13:16:34 +01:00
Sean Gilligan
42eb791d39
SeedPeers: use ByteBuffer
directly for address conversion
...
IPv4 addresses in `int` form don't care about signedness, and we
won't do math on them.
2023-03-24 11:31:26 +01:00
Andreas Schildbach
94bdbc29a0
UTXO, TransactionOutputChanges: remove unused serialization code
...
This was left over when we removed `UTXOsMessage` in d07e75e2a8
.
2023-03-24 10:49:32 +01:00
Andreas Schildbach
dadeb71cb3
VersionMessage: use skipBytes()
to skip over a value we don't ware about
2023-03-23 18:47:39 +01:00
Andreas Schildbach
ed57fa08ed
Message: make serializer
field immutable
...
This requires two new pass-through constructors, but in return we
get rid of the setter.
2023-03-23 13:34:31 +01:00
Andreas Schildbach
b8f77abe62
SeedPeers: fix conversion of int representation of IPv4 addresses to InetAddress
...
Usage of `writeUint32LE()` is wrong on two counts:
* `InetAddress.getByAddress()` expects big-endian values, and the method
above encodes as little-endian.
* Java int is signed, even if the Java hex literal appears unsigned. The
above method expects unsigned values.
So we add the previously missing `writeInt32BE()` to ByteUtils and use it.
2023-03-23 00:28:37 +01:00
Andreas Schildbach
72370d0ff6
Sha256Hash: add reading from and writing to ByteBuffer
...
Includes a test.
2023-03-22 22:19:54 +01:00
Andreas Schildbach
30dec4ff7c
Message: replace readBytes(1)[0]
with readByte()
2023-03-22 21:39:05 +01:00
Andreas Schildbach
89198853d1
TransactionInput: check sequence
bounds
...
It turns out a test is supplying an invalid fake value.
This is now fixed.
2023-03-22 20:32:31 +01:00
Andreas Schildbach
014939410e
TransactionOutPoint: check index
bounds
...
It turns out some tests are supplying an invalid fake value for
coinbase transaction. This is now fixed.
2023-03-22 20:23:56 +01:00
Andreas Schildbach
c577303aed
TransactionTest: don't use a custom protocol version in a test
...
The current default will do fine.
2023-03-22 20:06:41 +01:00
Andreas Schildbach
a7bd460865
Message: make params
field immutable
2023-03-22 18:46:11 +01:00
Andreas Schildbach
be12c389f3
Message: parse the original ByteBuffer
, rather than a copy of it
...
This has the advantage of using the read position contained in
`ByteBuffer`, so we remove our `offset` and `cursor` fields.
2023-03-22 15:04:05 +01:00
Andreas Schildbach
32837798c1
Message: remove length
and optimalEncodingMessageSize
fields
...
This does away with a lot of updating the fields while parsing and
mutating the data.
`getMessageSize()` now returns the length of the data as
serialized by `bitcoinSerializeToStream()`, which can be different
from the length of an original payload.
The default implementation of `getMessageSize()` is not performant
enough for all usecases, to it is overridden in strategic subclasses
to calculate the size in a much more efficient way.
2023-03-22 14:31:26 +01:00
Andreas Schildbach
7fba2d8c6b
Message: fold unsafeBitcoinSerialize()
into bitcoinSerialize()
2023-03-22 02:21:21 +01:00