TryFrom` became available in Rust 1.34 so we can use it now we have
bumped our MSRV.
Add a macro for implementing `TryFrom` for various lists of
`sha256::Hash` types. Use the macro to for vec, slice, and boxed slice.
Clippy emits:
warning: this function has an empty `#[must_use]` attribute, but
returns a type already marked as `#[must_use]`
This is because the return type of the function
`legacy_encode_signing_data_to` is `EncodeSigningDataResult` which is
already marked as `must_use`. There is no need to have `must_use` on the
function also.
I'm guessing this got through to master because we only just added
clippy to CI.
Clippy warns about creating a reference that is immediately
de-referenced.
Remove unnecessary explicit `&`, while we are at it remove unnecessary
explicit types that appear on the same lines of code.
57dd6739c3 Do not print error when displaying for std builds (Tobin C. Harding)
b80cfeed85 Bind to error_kind instead of e (Tobin C. Harding)
241ec72497 Bind to b instead of e (Tobin C. Harding)
01f481bf5c Bind to s instead of e (Tobin C. Harding)
5c6d369289 network: Remove unused error variants (Tobin C. Harding)
e67e97bb37 Put From impl below std::error::Error impl (Tobin C. Harding)
6ca98e5275 Remove error TODO (Tobin C. Harding)
Pull request description:
As part of the ongoing error improvement work and as a direct result of [this comment](https://github.com/rust-bitcoin/rust-bitcoin/pull/987#issuecomment-1135563287) improve the `Display` implementations of all our error types so as to not repeat the source error when printing.
The first 5 patches are trivial clean ups around the errors. Patch 6 is the real work.
EDIT: ~CC @Kixunil, have I got the right idea here bro?~ Patch 6 now includes a macro as suggested.
ACKs for top commit:
Kixunil:
ACK 57dd6739c3
apoelstra:
ACK 57dd6739c3
sanket1729:
ACK 57dd6739c3. Did not check if we covered all cases. We need to remember to use `write_err!` instead of `write!` in future.
Tree-SHA512: 1ed26b0cc5f9a0f71684c431cbb9f94404c116c9136be696434c56a2f56fd93cb5406b0955edbd0dc6f8612e77345c93fa70a70650118968cc58e680333a41de
99f565f932 Add non_exhaustive to all error enums (Tobin C. Harding)
Pull request description:
Adding an error variant to a public enum is an API breaking change, this means making, what could be, small refactorings or improvements harder. If we use `non_exhaustive` for error types then we mitigate this cost.
There is a tradeoff however, downstream users who explicitly match on our public error types must include a wildcard pattern.
ACKs for top commit:
apoelstra:
ACK 99f565f932
Kixunil:
ACK 99f565f932
Tree-SHA512: ff329f87d52b3fbe24654f32e4062ddae73173cba5a13d511591158e68ee278e9bdc0a70a3e0b42d6606b369255923f9c46d8b3d1b2ff75f8461a82567df80cd
Adding an error variant to a public enum is an API breaking change, this
means making what could be small refactorings or improvements harder. If
we use `non_exhaustive` for error types then we mitigate this cost.
There is a tradeoff however, downstream users who explicitly match on
our public error types must include a wildcard pattern.
97a5bb1439 Implement std::error::source codebase wide (Tobin C. Harding)
0a9191b429 Add parenthesis around left hand side of companion (Tobin C. Harding)
7cf8af2f86 Put Error impl block below Display (Tobin C. Harding)
2384712364 Re-order Display match arms (Tobin C. Harding)
Pull request description:
Now that we have MSRV of 1.41.1 we should use `source` instead of `cause`. Audit the whole codebase and implement `source` for _every_ error type we have.
The first three patches are preparatory cleanup, patch 3 is particularly shameful (adds parenthesis to make my editor work).
CC @Kixunil because he is championing the error stuff.
ACKs for top commit:
apoelstra:
ACK 97a5bb1439
Tree-SHA512: 46313a28929445f32e01e30ca3b0246b30bc9d5e43db5754d4b441e9c30d3e427efaf247100eb6b452f98beec5a4fcde1daba7943a772114aa34f78ab52cbc60
Improve documentation in `psbt/mod.rs` by doing:
- Use full sentences (full stops and capitalisation)
- Use 100 line column width
- Use back ticks and links as appropriate
- Use `Errors` section
- Use third person tense to describe functions
Improve the rustdoc documentation in the `sighash` module by doing:
- Improve grammar
- Use full sentences (full stops and capitalisation)
- Use 100 line column width
- Use back ticks and links as appropriate
- Improve correctness of `SigHashCache::new` function
Use cargo to upgrade from edition 2015 to edition 2018.
cargo fix --edition
No manual changes made. The result of the command above is just to fix
all the use statements (add `crate::`) and fix the fully qualified path
formats i.e., `::Foo` -> `crate::Foo`.
Recently we added a bunch of additional sighash types, some of the code
comments became stale. Use the non-specific term 'sighash type' instead
of a particular sighash identifier in comments to make the comments more
applicable.
Recently we update all types and docs to use `Sighash` instead of
`SigHash` because 'sighash' is a single word. We should apply the same
logic to functions and variable names.
Do not use an underscore in the identifier 'sighash'.
Our usage of `SigHash` implies that 'sighash' is _two_ words; 'sighash'
is a well known word in the Bitcoin ecosystem it should appear in
identifiers as `Sighash`.
Rename the `SigHash` type to `Sighash`.
Our usage of `SigHash` implies that 'sighash' is _two_ words; 'sighash'
is a well known word in the Bitcoin ecosystem it should appear in
identifiers as `Sighash`.
Rename the `InvalidSigHashType` variant to `InvalidSighashType`.
Our usage of `SigHash` implies that 'sighash' is _two_ words; 'sighash'
is a well known word in the Bitcoin ecosystem it should appear in
identifiers as `Sighash`.
Rename `SigHashCache` to `SighashCache`.
Our usage of `SigHash` implies that 'sighash' is _two_ words; 'sighash'
is a well known word in the Bitcoin ecosystem it should appear in
identifiers as `Sighash`.
Rename `SigHashTypeParseError` to `SighashTypeParseError`.
Our usage of `SigHash` implies that 'sighash' is _two_ words; 'sighash'
is a well known word in the Bitcoin ecosystem it should appear in
identifiers as `Sighash`.
Rename `SchnorrSigHashType` to `SchnorrSighashType`.
Our usage of `SigHash` implies that 'sighash' is _two_ words; 'sighash'
is a well known word in the Bitcoin ecosystem it should appear in
identifiers as `Sighash`.
Rename `EcdsaSigHashType` to `EcdsaSighashType`.
992857ad0a PsbtSighashType unit tests (Dr Maxim Orlovsky)
5be1cdb8c7 PsbtSigHashType Display and FromStr implementation (Dr Maxim Orlovsky)
7cdcdaad6c Support SIGHASH_RESERVED in SchnorrSigHashType::from_u8 (Dr Maxim Orlovsky)
Pull request description:
The newly introduced `PsbtSigHashType` uses very different serde formatting from previously used `EcdsaSigHashType`; for instance it does not output human-readable sighash. This is especially obvious when printing out PSBT as JSON/YAML object and is a breaking change from the `0.27`. Serde human-readable implementation requires `Display/FromStr`, which were also absent.
ACKs for top commit:
sanket1729:
ACK 992857ad0a. This is much better
apoelstra:
ACK 992857ad0a
Tree-SHA512: 71a46471f34b5481e4c1273a66846f59d61bfd98fcb65e7823ca216ff0dd419d81ca86d99c7aaf674fcfe2b1c010e899c8e74328f60a1e809015c663c453cc89
Rust naming conventions stipulate that conversion methods from owned ->
owned for `Copy` types use the naming convention `to_`.
This change makes the function name objectively better, however it makes
no claims of being the 'best' name. We have had much discussion on using
`to_standard` vs `to_u32` but are unable to reach consensus.
We currently implement `Display` and `FromStr` on `EcdsaSigHashType` and
use them in the `serde_string_impl` macro to implement ser/de.
Mirror this logic in `SchnorrSigHashType`.
The exact code formatting we use is not as important as uniformity.
Since we do not use tooling to control the formatting we have to be
vigilant ourselves. Recently I (Tobin) changed the way default type
parameters were formatted (arbitrarily but uniformly). Turns out I
picked the wrong way, there is already a convention as shown in the rust
documentation online (e.g. [1]).
Use 'conventional' spacing for default type parameters. Make the change
across the whole repository, found using
git grep '\<.* = .*\>'
[1] - https://doc.rust-lang.org/book/ch19-03-advanced-traits.html
This avoids some allocation of creating a vec of TxOut to
create a slice incase the data is already available in psbt/other
methods. Facilitates creation of Prevouts from &[TxOut] as well as
&[&TxOut]
Rust convention is to use `to_` for conversion methods that convert from
an owned type to an owned `Copy` type. `into_` is for owned to owned
non-`Copy` types.
Re-name conversion methods that use `into_` for `Copy` types to use
`to_`, no need to deprecate these ones because they are unreleased.
Rust idiomatic style is to put the rustdoc _above_ any attributes on
types, functions, etc.
Audit the codebase and move comments/attributes to the correct place.
Add a trailing full stop at times to neaten things up a little extra.
7d982fa9a2 Add all tests from BIP 371 (sanket1729)
d22e0149ad Taproot psbt impl BIP 371 (sanket1729)
108fc3d4db Impl encodable traits for TapLeafhash (sanket1729)
c7478d8fd0 Derive serde for taproot stuctures (sanket1729)
Pull request description:
Built on top of #677 . Will rebase and mark ready for review after #677 is merged.
ACKs for top commit:
apoelstra:
ACK 7d982fa9a2
dr-orlovsky:
re-tACK 7d982fa9a2 basing on `git range-diff`. The original PR before last re-base was tested commit-by-commit.
Tree-SHA512: feb30e4b38d13110a9c0fabf6466d8f0fb7df09a82f4e01d70b8371b34ab0187004a6c63f9796c6585ee30841e8ee765ae9becae139d2e1e3d839553d64c3d1e
106acdc3ac Add fuzzing for Witness struct (Riccardo Casatta)
2fd0125bfa Introduce Witness struct mainly to improve ser/de performance while keeping most usability. (Riccardo Casatta)
Pull request description:
At the moment the Witness struct is `Vec<Vec<u8>>`, the vec inside a vec cause a lot of allocations, specifically:
- empty witness -> 1 allocation, while an empty vec doesn't allocate, the outer vec is not empty
- witness with n elements -> n+1 allocations
The proposed Witness struct contains the serialized format of the witness. This reduces the allocations to:
- empty witness -> 0 allocations
- witness with n elements -> 1 allocation for most common cases (you don't know how many bytes is long the entire witness beforehand, thus you need to estimate a good value, not too big to avoid wasting space and not too low to avoid vector reallocation, I used 128 since it covers about 80% of cases on mainnet)
The inconvenience is having slightly less comfortable access to the witness, but the iterator is efficient (no allocations) and you can always collect the iteration to have a Vec of slices. If you collect the iteration you end up doing allocation anyway, but the rationale is that it is an operation you need to do rarely while ser/de is done much more often.
I had to add a bigger block to better see the improvement (ae860247e191e2136d7c87382f78c96e0908d700), these are the results of the benches on my machine:
```
RCasatta/master_with_block
test blockdata::block::benches::bench_block_deserialize ... bench: 5,496,821 ns/iter (+/- 298,859)
test blockdata::block::benches::bench_block_serialize ... bench: 437,389 ns/iter (+/- 31,576)
test blockdata::block::benches::bench_block_serialize_logic ... bench: 108,759 ns/iter (+/- 5,807)
test blockdata::transaction::benches::bench_transaction_deserialize ... bench: 670 ns/iter (+/- 49)
test blockdata::transaction::benches::bench_transaction_get_size ... bench: 7 ns/iter (+/- 0)
test blockdata::transaction::benches::bench_transaction_serialize ... bench: 51 ns/iter (+/- 5)
test blockdata::transaction::benches::bench_transaction_serialize_logic ... bench: 13 ns/iter (+/- 0)
branch witness_with_block (this one)
test blockdata::block::benches::bench_block_deserialize ... bench: 4,302,788 ns/iter (+/- 424,806)
test blockdata::block::benches::bench_block_serialize ... bench: 366,493 ns/iter (+/- 42,216)
test blockdata::block::benches::bench_block_serialize_logic ... bench: 84,646 ns/iter (+/- 7,366)
test blockdata::transaction::benches::bench_transaction_deserialize ... bench: 648 ns/iter (+/- 77)
test blockdata::transaction::benches::bench_transaction_get_size ... bench: 7 ns/iter (+/- 0)
test blockdata::transaction::benches::bench_transaction_serialize ... bench: 50 ns/iter (+/- 5)
test blockdata::transaction::benches::bench_transaction_serialize_logic ... bench: 14 ns/iter (+/- 0)
```
With an increased performance to deserialize a block of about 21% and to serialize a block of about 16% (seems even higher than expected, need to do more tests to confirm, I'll appreciate tests results from reviewers)
ACKs for top commit:
apoelstra:
ACK 106acdc3ac
sanket1729:
ACK 106acdc3ac
dr-orlovsky:
utACK 106acdc3ac
Tree-SHA512: e4f23bdd55075c7ea788bc55846fd9e30f9cb76d5847cb259bddbf72523857715b0d4dbac505be3dfb9d4b1bcae289384ab39885b4887e188f8f1c06caf4049a
Witness struct is in place of the Vec<Vec<u8>> we have before this commit.
from_vec() and to_vec() methods are provided to switch between this type and Vec<Vec<u8>>
Moreover, implementation of Default, Iterator and others allows to have similar behaviour but
using a single Vec prevent many allocations during deserialization which in turns results in
better performance, even 20% better perfomance on recent block.
last() and second_to_last() allows to access respective element without going through costly Vec
transformation
Instead of always requiring the full raw script and leaf version, allow
just specifying a raw leaf hash to the sighash computation functions.
This is very useful when dealing with PSBTs, because the
`PSBT_IN_TAP_BIP32_DERIVATION` field only maps a public key to a leaf
hash, so a signer could just take it and produce a signature with it
rathern than having to jump through hoops to recover the full raw
script.
This is the initial step towards using and maybe enforcing clippy.
It does not fix all lints as some are not applicable. They may be
explicitly ignored later.
Tobin C. Harding
Andrew Poelstra
Dawid Ciężarkiewicz
sanket1729
eunoia_1729
Dr Maxim Orlovsky
KaFai Choi
Riccardo Casatta
Alekos Filini
Martin Habovstiak