Moves all of the content from `bitcoin/p2p` into `p2p`.
`TryFrom<Network>` must be implemented for `Network -> Magic` now that
`Network` is a foreign, non-exhaustive type. Ser/de test is updated
accordingly, as well as the `Magic::from_network` constructor, which I
have opted to return an `Option<Self>`. The `TryFrom` implementation
uses a new `UnknownNetworkError(Network)` that mirrors the `Magic ->
Network` error.
The example handshake does not generate a random nonce for the version
message, as there is no `rand` dependency in this crate and the program
only makes a single, user-defined connection.
It appears we can do better than copying and pasting the consensus
encoding macros and functions from `bitcoin` without doing weird
cross-crate macros that require some special knowledge of the crate to
know when they will compile.
All of the `Encodable` implementations are defined within `p2p` with the
exception of `p2p` types that are a `Vec`. To fully decouple `p2p` from
`encode` we can define these in `p2p` like the others. This is the final
step in removing anything `p2p` related from the rest of `bitcoin`.
In preparation to move `p2p` to its own crate, any `Vec<T>` that satisfies
`T: Encodable` will not trivially implement `Encodable` because of the
orphan rule. A type defined within p2p may implement `Encodable`,
however, and the simplest possible type is one that simply wraps the
vector. Three types that will implement `Encodable` within `p2p` are
added here.
d6296cd3d1 Remove usage of hex::test_hex_unwrap (Tobin C. Harding)
37035e20e8 Simplify and improve transaction benchmarks (Tobin C. Harding)
Pull request description:
We have the `hex_lit` dependency for converting a hex string literal to an array.
Currently we have a `test_hex_unwrap` macro in the `hex v0.3.0` release but not on either `master`
or the upcoming `v1.0.0-alpha.0` release. This is making PRs around releasing and depending on the
release more noisy than required.
Introduce a `test_hex_unwrap` macro in internals for usage when the input is not a string literal.
Use `hex_lit::hex` where possible (often needing an additional call to
ACKs for top commit:
apoelstra:
ACK d6296cd3d1989cf28d67a5329ad60da4f814ba92; successfully ran local tests
Kixunil:
ACK d6296cd3d1
Tree-SHA512: eab3573f6b7fee408ae11821b77e56cbaddf7cc4540bdc31ed7ef9eb3f25987f50e484f1553aaaa9709367e614eb77ed36250875d0faf5a51ab3fe709d4d4054
We have the `hex_lit` dependency for converting a hex string literal
to an array.
Currently we have a `test_hex_unwrap` macro in the `hex v0.3.0` release
but not on either `master` or the upcoming `v1.0.0-alpha.0` release.
This is making PRs around releasing and depending on the release more
noisy than required.
Use `hex_lit::hex` where possible (often needing an additional call to
`to_vec()`) and where not possible use `Vec::from_hex`.
The `hash_newtype` macro is explicitly designed to produce a hash that
is not a general purpose hash type to try and prevent users hashing
arbitrary stuff with it. E.g., `Txid` isn't meant to be just hash
arbitrary data. However we provide a `From` impl that will convert any
instance of the inner hash type into the new type. This kind of defeats
the purpose. We provide `from_byte_array` and `to_byte_array` to allow
folk to 'cast' from one hash type to another if they really want to and
its ugly on purpose.
Also, it is becoming apparent that we may be able to remove the `hashes`
crate from the public API of `primitives` allowing us to stabalise
`primitives` without stabalising `hashes`.
For both these reasons remove the `From` impl from the `hash_newtype`
macro. Note that deprecating doesn't seem to work so we just delete it.
This function is deprecated, stop using it in favour of
`Hash::from_byte_array`.
Patch only touches test code, I'm guessing that is why lint warnings
were no showing up.
BIP324 introduced the second version of p2p network messages which
are designed to be used with the new encrypted transport. This patch
adds a V2NetworkMessage type which wraps a NetworkMessage and handles
the V2 encoding and decoding. It is a parallel of the existing
RawNetworkMessage type (which may be better described as
V1NetworkMessage). A priority of this patch was to not re-invent any
wheels and try to use the existing patterns as much as possible.
While the hash value of the Error variant is meaningless, the variant
still conforms to all other Inventory messages and requires a 32
byte hash to be sent over the wire. This is how bitcoin core operates.
This patch adds the 32 byte array to the Error variant in order to make
its Encoding and Decoding paths symmetrical. This also allows a reader
to discard the 32 bytes when decoding a message. The hash is still not
exposed to the caller.
This was never a problem before because the top level RawNetworkPackage
pulls all the required bytes off a reader before decoding. But this is
not as easy to do with the v2 p2p network messages.
There is a range of different wordings used in the docs of constructor
type functions.
Change all to start with `Constructs a new` or `Constructs an empty`.
In functions that act like constructors there is a mixture of the usage
of `creates` and `constructs`.
Replace all occurrences of `creates` with `constructs` in the first line
of docs of constructor like functions.
The `encode::Error::ParseFailed` variant holds an inner string, this is
suboptimal.
In an effort to patch the `encode::Error` while mimizing the diffs
required add a helper function that creates the variant. The benefit is
that later patches that effect this variant will only need to update the
constructor function and not every call site.
Internal change only.
323e706113 Add rustfmt config option style_edition (Tobin C. Harding)
2e4179ed0f Run the formatter (Tobin C. Harding)
2c40b4f4ec Configure formmater to skip read_compact_size (Tobin C. Harding)
Pull request description:
`rustfmt` is emitting:
Warning: the `version` option is deprecated. Use `style_edition` instead.
As suggested add a config option and set it to 2021.
- Patch 1: Manually configure rustfmt to skip some code
- Patch 2: Run the formmater with current configuration
- Patch 3: Add the new config option (remove old one), introduces no new formatting requirements
ACKs for top commit:
apoelstra:
ACK 323e706113 successfully ran local tests
Tree-SHA512: 7f80cc89f86d2d50936e51704344955fa00532424c29c0ee3fae1a6836e24030f909b770d28da13e1c5efde3d49ad7d52c6d909d120fb09c33abf1755f62cd38
If folk really want to index into a hash they can us `as_byte_array`
then index that.
Includes a bump to the version number of `hashes` to `v0.15.0` - this
is because otherwise `secp` won't build since we are breaking an API
that is used in the current release of secp.
Fix: #3115
At some stage we named the compact encoding `VarInt` (which makes sense
because the compact size encoding is a variable length integer encoding).
However it turns out the term "varint" is used in Core for a different
encoding so this may lead to confusion.
While we fix this naming thing observe also that the `VarInt` type is
unnecessarily complicated, all we need to be able to do is encode and
decode integers in compact form as specified by Core. We can do this
simply by extending our `WriteExt` and `ReadExt` traits.
Add `emit_compact_size` and `read_compact_size` to emit and read compact
endcodings respectively.
Includes addition of `internals::compact_size::encoded_size_const`.
Patch originally written by Steven, Tobin cherry-picked and did a bunch
of impovements after the varint vs compact_size thing (#1016).
ref: https://en.bitcoin.it/wiki/Protocol_documentation#Variable_length_integer
Co-developed-by: Tobin C. Harding <me@tobin.cc>
We already explicitly do not support 16 bit machines.
Also, because Rust supports `u182`s one cannot infallibly convert from a
`usize` to a `u64`. This is unergonomic and results in a ton of casts.
We can instead limit our code to running only on machines where `usize`
is less that or equal to 64 bits then the infallible conversion is
possible.
Since 128 bit machines are not a thing yet this does not in reality
introduce any limitations on the library.
Add a "private" trait to the `internals` crate to do infallible
conversion to a `u64` from `usize`.
Implement it for all unsigned integers smaller than `u64` as well so
we have the option to use the trait instead of `u32::from(foo)`.
Wildcards have been replaced with what is actually used.
In a couple of cases an additional use statement was added to the test
module to import `DisplayHex` which is only used in test, but
previously imported with the wildcard at the top.
the `blockdata` directory is code organisation thing, all the
types/modules are re-exported from other places. In preparation for, and
to make easier, the `primitives` crate smashing work - remove all
explicit usage of `blockdata`.
Note that the few instances remain as they seem required e.g.,
`pub(in crate::blockdata::script)`
Refactor only, no logic changes.
Currently we have a trait `Hash` that is required for `Hmac`, `Hkdf`,
and other use cases. However, it is unegonomic for users who just want
to do a simple hash to have to import the trait.
Add inherent functions to all hash types including those created with
the new wrapper type macros.
This patch introduces some duplicate code but we are trying to make
progress in the hashes API re-write. We can come back and de-dublicate
later.
Includes making `to_byte_array`,`from_byte_array`, `as_byte_array`, and
`all_zeros` const where easily possible.
I'd like to take out the `payload` of RawNetworkMessage and
then send it to the actual network message processor, but
finds there is no way to do it. This commit adds such an API
to expose the owned value of inner `payload`.
We just added to now types that are thin wrappers around `u32`s for
block heights and intervals.
Add `Encodable` and `Decodable` impls and use the new types. While we
are at it re-export the types from the crate root so users don't have to
dig into the `units` crate.
Clean up the test imports in the `p2p` module:
- Use `use super::*` as is conventional.
- Use `sha256d::Hash` as is conventional.
Refactor, no logic changes.
6ddb5cce37 Use Magic::BITCOIN in unit tests (Tobin C. Harding)
Pull request description:
We are currently calling `From` to create the magic bytes, this is unnecessary since `Magic` provides consts.
Refactor only, no logic changes.
ACKs for top commit:
Kixunil:
ACK 6ddb5cce37
apoelstra:
ACK 6ddb5cce37
Tree-SHA512: 20e2e017683f123309e3c0876bba42d86a9411bb225f07c486716184fc79837e04a832338ec8b18874ac76791260f6a4620b932ede92c8b222dac08d468cef8a
Our decoding code reads bytes in very small chunks. Which is not
efficient when dealing with the OS where the cost of a context switch is
significant. People could already buffer the data but it's easy to
forget it by accident.
This change requires the new `io::BufRead` trait instead of `io::Read`
in all bounds.
Code such as `Transaction::consensus_decode(&mut File::open(foo))` will
break after this is applied, uncovering the inefficiency.
This was originally Kix's work, done before we had the `io` crate.
Changes to `bitcoin` were originally his, any new mistakes are my own.
Changes to `io` are mine.
Co-developed-by: Martin Habovstiak <martin.habovstiak@gmail.com>
There is no advantage in having `io::Read` as opposed to `Read` and
importing the trait. It is surprising that we do so.
Remove `io::` path from `io::Read` and `io::Write`. Some docs keep the
path, leave them as is. Add import `use io::{Read, Write}`.
Refactor only, no logic changes.
We have a convention in `rust-bitcoin` to use external crates directly
when importing them not via `crate::foo`.
Update all the import paths for `io` to use this form.
In order to move towards our own I/O traits in the `rust-bitcoin`
ecosystem, we have to slowly replace our use of the `std` and
`core2` traits.
Here we take the second big step, replacing
`{std,core2}::io::Read` with our own `bitcoin_io::io::Read`. We
provide a blanket impl for our trait for all `std::io::Read`, if
the `std` feature is enabled, allowing users who use their own
streams or `std` streams to call `rust-bitcoin` methods directly.
8eff4d0385 Remove private hex test macro (Tobin C. Harding)
Pull request description:
We have this macro in `hex-conservative` now, remove the version here.
This patch does not change the public API and only touches test code.
ACKs for top commit:
apoelstra:
ACK 8eff4d0385
clarkmoody:
ACK 8eff4d0385
Tree-SHA512: 93a08fff778930071cd1a28c19202e4a94ca8881b2e873538de2e942b71c2cd6184ed6364c572538a8a699295a71761c6f836accaf251a15683138b71f148fab
On our way to v1.0.0 we are defining a standard for our error types,
this includes:
- Uses the following derives (unless not possible, usually because of `io::Error`)
`#[derive(Debug, Clone, PartialEq, Eq)]`
- Has `non_exhaustive` unless we really know we can commit to not adding
anything.
Furthermore, we are trying to make the codebase easy to read. Error code
is write-once-read-many (well it should be) so if we make all the error
code super uniform the users can flick to an error and quickly see what
it includes. In an effort to achieve this I have made up a style and
over recent times have change much of the error code to that new style,
this PR audits _all_ error types in the code base and enforces the
style, specifically:
- Is layed out: definition, [impl block], Display impl, error::Error impl, From impls
- `error::Error` impl matches on enum even if it returns `None` for all variants
- Display/Error impls import enum variants locally
- match uses *self and `ref e`
- error::Error variants that return `Some` come first, `None` after
Re: non_exhaustive
To make dev and review easier I have added `non_exhaustive` to _every_
error type. We can then remove it error by error as we see fit. This is
because it takes a bit of thinking to do and review where as this patch
should not take much brain power to review.
We would like the codebase to be optimized for readability not ease of
development, as such code that is write-once-read-many should not use
macros.
Currently we use the `impl_std_error` macro to implement
`std::error::Error` for struct error types. This makes the code harder
to read at a glance because one has to think what the macro does.
Remove the `impl_std_error` macro and write the code explicitly.
Throughout the codebase we cast values to `u64` when constructing a
`VarInt`. We can make the code marginally cleaner by adding `From<T>`
impls for all unsigned integer types less than or equal to 64 bits.
Also allows us to (possibly unnecessarily) comment the cast in a single
place.
rustaceanrob
Fmt Bot
merge-script
Bilog WEB3
Tobin C. Harding
kilavvy
Nick Johnson
yancy
Jamil Lambert, PhD
Steven Roose
leichak
Liu-Cheng Xu
Matt Corallo