Currently we copy data into a new buffer before passing it into the
hasher, we can just hash the data directly.
Internal change only, no external change.
Fix: #2917
4576f4ba69 Remove fuzz rust version (yancy)
Pull request description:
Fuzz is not compatible with the workspace MSRV.
Closes https://github.com/rust-bitcoin/rust-bitcoin/issues/2923
ACKs for top commit:
tcharding:
ACK 4576f4ba69
apoelstra:
ACK 4576f4ba69
Kixunil:
ACK 4576f4ba69
Tree-SHA512: 3b350e1f22a585719eb272de41b6a2b7a83a8d7810df8d98c4ecc06cf391ec3c3582934abc05d53bbdbe4ae4165fead9df2abf6f5aa84a74c1c0a7a7569eac6c
60aa4acc5f Update README instructions for running MSRV (yancy)
Pull request description:
The file mentioned in the README for running MSRV no longer exists
closes https://github.com/rust-bitcoin/rust-bitcoin/issues/2934
ACKs for top commit:
tcharding:
ACK 60aa4acc5f
apoelstra:
ACK 60aa4acc5f Definite improvement. Unfortunately will not work for downstream users but currently we do not have a solution for them I think
Kixunil:
ACK 60aa4acc5f
Tree-SHA512: 554b1b4d6ea8a929fb168e48b8925e20e6ee717a603a4008793376393d1ae8ba18bc290fb69747be155cd75439499500056c54e845a6fa6dc097e523b76ea54f
`fmt_value_in` was added when `display_in` wasn't available. However
common usage patterns seem to favor `display_in`. It can be used within
format strings and supports formatting options.
Removing it will simplify the codebase, so this deprecates it.
Introduce an empty `bitcoin-primitives` crate.
We were give the name on crates.io and previously a version
`v0.1.16-alpha` was released so we use `v0.100.0`.
15f6bacec9 api: Run just check-api (Ryan Breen)
9684d496bb Add is_standard_op_return (Ryan Breen)
Pull request description:
This is the suggestion for #2292 to check OP_RETURN length
ACKs for top commit:
apoelstra:
ACK 15f6bacec9
Tree-SHA512: e346b5eff7cc40b98a08948c83cb5c064184541d819c37a977e432ec09df7f9e1a074f16a4df598142784bd875f1379e2b0848fe898923e4e12829f85b4c4520
Low precision was previously not considered because it was believed it
could lead to misleading data being displayed. However it's quite
possible that people using low precision know what they are doing and
it's sometimes useful to show rounded numbers.
To enable low precision we just compute what the rounded number would be
and display that one instead.
This actually fully closes#2136 since this issue was mentioned there
along with previously-fixed displaying of higher precisions.
Displaying with minimal number of zeros is the default in Rust and we
want to follow it. This was originally implemented in #716 but #2604
reversed it claiming this is common, however it broke people who rely on
minimal display (e.g. BIP21) without fixing the root cause of #2136.
This reverts commit d887423efc adds a test
to prevent this change and also fixes the problem with `{:0.8}` not
working.
Closes#2136Closes#2948
7f024c333e Revert "bug fix" (actually a bug) from #2655 (Martin Habovstiak)
Pull request description:
In a2b019f82 it was claimed that losing the input string without alloc is a bug. It is not, because allowing no-alloc is desired and there's no way to have the input string otherwise so we just accept it'll be missing and modify the messages accordingly. The commit that forced alloc was also horribly misleading since it kept the `alloc` feature but it makes this one easier.
Note that input string is still present by default in all configurations except for no-alloc.
I think we should also backport this and release fixed `units` because of the misleading `alloc` feature in them. Although it's not urgent. The only crate I know of that is kinda broken by it is `ln_types` which is already broken by depending on old `bitcoin`.
ACKs for top commit:
apoelstra:
ACK 7f024c333e
tcharding:
ACK 7f024c333e
Tree-SHA512: 014ed823f0daf2c47ca6cedf1aad0d94b702f2ca53b096556a76566baeadb209b9d4d710872c2b8308542fd7cfe6d815a206f1a84174458d251bf30882be7719
e8a30bfca7 taproot: Lowercase error messages (Jamil Lambert, PhD)
717da4bf69 Capitalize Huffman (Jamil Lambert, PhD)
175f69abeb Capitalize Merkle (Jamil Lambert, PhD)
be5b93ea4f Do not capitalize tree (Jamil Lambert, PhD)
b48ca050f9 Remove tap tree (Jamil Lambert, PhD)
c8e6c13608 Capitalize Taproot (Jamil Lambert, PhD)
Pull request description:
Taproot, Merkle and Huffman are proper nouns and should be capitalized in strings and docs.
These have been capitalized everywhere.
All cases of `taproot`/`taptree`/`tap tree` in docs and strings have been changed to be `Taproot tree` or ```[`TapTree`]``` as appropriate.
All cases of `tree` have been changed to lower cases, except where it is in title case.
Error messages in `Taproot` have been changed to start with a lowercase character, except when it is a proper noun, and not end with a period.
Close: #2913
ACKs for top commit:
tcharding:
ACK e8a30bfca7
apoelstra:
ACK e8a30bfca7
Tree-SHA512: 7d06b717072f7894093df8501752374637fc6673bd7b6f1b95297ab076725eb0d7f48caa45c321fd13a45d0e4387ab9eeb398261f7988d3c9f90cfa75aee17ca
de8a8ca077 Automated update to Github CI to rustc nightly-2024-06-30 (Update Nightly Rustc Bot)
Pull request description:
Automated update to Github CI workflow `rust.yml` by [create-pull-request](https://github.com/peter-evans/create-pull-request) GitHub action
ACKs for top commit:
tcharding:
ACK de8a8ca077
Tree-SHA512: 7d5918439152919c18875b717bff7a262436efac034f2a1fa3562da0621c3286f8a4bc38952bf41333fe7f851b759acc43e7036e3044872e277470907579c08b
0630457403 Create empty bitcoin-addresses crate (Tobin C. Harding)
Pull request description:
We intend on splitting the address types and logic out into a separate crate. In preparation for doing so, and so that we can grab the name on crates.io, add an empty crate `bitcoin-addresses`.
Tie it in to the CI infrastructure excluding the `check-api` infrastructure because that is in flux.
ACKs for top commit:
Kixunil:
ACK 0630457403
apoelstra:
ACK 0630457403 will try to tag and publish tonight, but may be tomorrow
Tree-SHA512: faf54f43954a799f1003dae67930edc7f306c35403ba05e511df6685c1068850c47fe1a46afdcabcb32c3f43398803181c7f53d6bcf64a4aec7123e9ce9306e8
865ba3fc39 Move serde string macros to internals (Tobin C. Harding)
4a2b13fcde internals: Feature gate whole serde module (Tobin C. Harding)
Pull request description:
The macros are internal things and can live in `internals`. This will help with future crate smashing.
ACKs for top commit:
apoelstra:
ACK 865ba3fc39
Kixunil:
ACK 865ba3fc39
Tree-SHA512: 7b3f029206c690ecf2894e0ad099d391312f7f8ec65ac9b5d4d9f25e6827f92075dcc851d0940a0faf1e27e7d0a305b575c8cc790939b3f222d7a2920d4d24fe
c717f7f424 Improve docs on private Witness fields (Tobin C. Harding)
Pull request description:
The `Witness` type is a reasonable complex data structure, make an effort to clarify its structure in the docs on the private fields.
Private docs only.
(Original idea pulled out of #2133.)
ACKs for top commit:
Kixunil:
ACK c717f7f424
apoelstra:
ACK c717f7f424 much clearer, thanks!
Tree-SHA512: 9d54b7eeefec97e584fb5f275049dbac0473c949fae8ab05c6961d6fc424c17a058af7037c2220ef1446af294d78c68bfee741cfeca1b18ecc402935d8069dab
In a2b019f82 it was claimed that losing the input string without alloc
is a bug. It is not, because allowing no-alloc is desired and there's
no way to have the input string otherwise so we just accept it'll be
missing and modify the messages accordingly. The commit that forced
alloc was also horribly misleading since it kept the `alloc` feature but
it makes this one easier.
Note that input string is still present by default in all configurations
except for no-alloc.
d099b9c195 Remove wildcard from prelude import (Jamil Lambert, PhD)
Pull request description:
This patch replaces `prelude::*` wildcard imports with the types actually used. In a couple of cases `DisplayHex` was previously imported by the wildcard but was only used in the test module, an additional import was added to the test module instead of at the top where it causes an unused import warning.
Close: #2875
ACKs for top commit:
Kixunil:
ACK d099b9c195
tcharding:
ACK d099b9c195
Tree-SHA512: d59dfac0961d2649d509039a11c1b5574d81d05fef567a624cf15be2f587de796ea960ba5a08bef788199331c2f790fb06f7b393182538c7d8b1891ded119efc
The `Witness` type is a reasonable complex data structure, make an
effort to clarify its structure in the docs on the private fields.
Private docs only.
010141ecc9 api: Run just check-api (Tobin C. Harding)
cf800a1b07 Remove bech32 dependency from blockdata (Tobin C. Harding)
Pull request description:
We have a single usage of the `bech32` crate inside the `blockdata` module, to convert a `WitnessVersion` to a `Fe32`. We then have a single call site where we use the conversion in the `address` module.
This code was written without thinking to hard about the introduced dependency on `bech32`, in hindsite it shouldn't have been added.
In preparation for splitting a bunch of code in `blockdata` out into the `primitives` crate remove the `bech32` stuff from the `witness_version` module.
ACKs for top commit:
Kixunil:
ACK 010141ecc9
apoelstra:
ACK 010141ecc9
Tree-SHA512: 2d368ebf64ab7197b421e0dec48623f3fb03243081a988ff9ed2070135c8741efe24520c2aab496d54c52595808f10ee39816eaab8e3e4a7e64955cd25285670
a7731b2f33 api: Run just check-api (Marko Bencun)
47cba7a655 bip32: add from_32_slice method to DerivationPath (Marko Bencun)
Pull request description:
ChildNumber already has a `From<u32>` impl, but converting `&[u32]` to a `DerivationPath` was still difficult.
ACKs for top commit:
Kixunil:
ACK a7731b2f33
tcharding:
ACK a7731b2f33
Tree-SHA512: 2db94ee035e686102b8201f637422bb96bd79858aeffdb007594d722902d31a20f27e61b88fae8c854c80f785d9e7837b0158a046639ff8cc2d20d8883391842
dc10a49876 api: Run just check-api (Tobin C. Harding)
5e8f204581 Pass sigs and associated types by value (Tobin C. Harding)
Pull request description:
We should pass `Copy` types by value not by reference.
Currently this is not done in secp, but lets do it here in bitcoin.
Pass by value:
- `SerializedSignature`
- bitcoin sigs
- secp sigs
- secp `Message`
This is a continuation of the work to split up #2404 into manageable PRs.
ACKs for top commit:
apoelstra:
ACK dc10a49876
Kixunil:
ACK dc10a49876
Tree-SHA512: 8736eba067c74edb951c92357f5b3d0fc99c4fa6dc3beea579c10b3150873b74e8ec46c2c01f18818b37fca6e77c6b6edddeb6340edde6a9d8c28a4e69164c8c
ebb862cbaf Automated update to Github CI to rustc nightly-2024-06-26 (Update Nightly Rustc Bot)
Pull request description:
Automated update to Github CI workflow `rust.yml` by [create-pull-request](https://github.com/peter-evans/create-pull-request) GitHub action
ACKs for top commit:
tcharding:
ACK ebb862cbaf
Tree-SHA512: ec7ba43d784ff6245abda82ab7c27b247d1bfc654335fe384cb0990a4b36d752467282990452f485c911f996a95f8a7c163fe44a177ef1bd4cdcffc6f0e5ed06
We intend on splitting the address types and logic out into a separate
crate. In preparation for doing so, and so that we can grab the name on
crates.io, add an empty crate `bitcoin-addresses`.
Tie it in to the CI infrastructure.
6a7f780018 api: Run just check-api (Tobin C. Harding)
0c9223ac05 Manually format write_err statement (Tobin C. Harding)
43d7c750cc taproot: Add error types (Tobin C. Harding)
afe41c8a39 taproot: Split errors up (Tobin C. Harding)
Pull request description:
Currently there are a couple of errors in the `taproot` module that are too general, resulting in functions that return a general error type when a specific one would do.
Split two errors out and use them for for enum variants and function returns as possible.
Done as part of #2883
ACKs for top commit:
Kixunil:
ACK 6a7f780018
apoelstra:
ACK 6a7f780018
Tree-SHA512: bf5ed50dd8f913280d007f03124c7918c4b6cd642e67726bc3ffff23d9897f764a4391e167115668c3a1e951197485eba280fdbb8a0ce1bb7efb051388d13997
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.
We have a single usage of the `bech32` crate inside the `blockdata`
module, to convert a `WitnessVersion` to a `Fe32`. We then have a single
call site where we use the conversion in the `address` module.
This code was written without thinking to hard about the introduced
dependency on `bech32`, in hindsite it shouldn't have been added.
In preparation for splitting a bunch of code in `blockdata` out into the
`primitives` crate remove the `bech32` stuff from the `witness_version`
module.
10440b36a2 api: Run just check-api (Shing Him Ng)
452a7cc091 Re-export public functions, enums, and structs from units::parse in the bitcoin crate root (Shing Him Ng)
Pull request description:
Re-export UnprefixedHexError in the bitcoin crate root
Fixes#2874
ACKs for top commit:
tcharding:
ACK 10440b36a2
Tree-SHA512: 78a0676aa6d6fe63d810eeeebd3bf053d679d966d0aebf752760bfb440b2d4e831b0ff6b54b0306b390126b3f96a70c0846b4a1ec20aca96994d4a7b75e438e3
31a2324bfd Fix fuzz script (Tobin C. Harding)
Pull request description:
Currently the `fuzz.sh` script fails with "unbound variable" if called without any arguments, this has gone undetected since we added `set -euox pipefail` because in CI we always call it with an argument.
Use chatGPT to fix the bug.
Fix: #2924
ACKs for top commit:
apoelstra:
ACK 31a2324bfd
Tree-SHA512: 28e018bafd02c4034863c84e3b491c83962379d41bddeae184c4e35b889cb3e5aed9605b01749a97c60e3090525db52786574e487ff20cbf61fbdc8eaa8ab5a0
8ee1744b9b Make 'use core::fmt' calls consistent (Shing Him Ng)
Pull request description:
I started taking a look at #2869 and looked for everything that was implementing the `Display` trait:
```rust
impl fmt::Display for _
```
but found some places where the imports weren't consistent:
```rust
impl Display for _
```
There were only a few instances of the latter, so I went ahead and cleaned those up before starting #2869
I started pulling this thread when I saw the same thing was happening for `fmt::Debug` and `fmt::Formatter` so I updated the rest of the `use core::fmt::*` statements with a few exceptions:
- No updates to `use core::fmt::*` if it was being called from within a function since I felt like the function scope was small enough to not cause confusion
- No updates to `use core::fmt::{self, Write as _};`
ACKs for top commit:
Kixunil:
ACK 8ee1744b9b
tcharding:
ACK 8ee1744b9b
Tree-SHA512: 33eb6ea0c4e808ef78bc87de6547144b756bde206c50d80488f740e97cd8d11f1abcb8936c487d7bfd29be5e21c7f40ff88f82acdaaec9aacb4b6362ffc4c680
bc25ed35d5 Order serde feature list alphabetically (Tobin C. Harding)
5bd3387c15 Move package metadata to be underneath package section (Tobin C. Harding)
a2a9f193fe Put workspace crates in alphabetical order (Tobin C. Harding)
05931cc0fa Run the formatter (Tobin C. Harding)
Pull request description:
We are getting an increasing number of crates in the repo, clean up the manifests a bit in an endevour to help keep things manageable.
All patches are trivial and the PR makes no logic changes.
ACKs for top commit:
Kixunil:
ACK bc25ed35d5
apoelstra:
ACK bc25ed35d5
Tree-SHA512: a9850449a6f71ac5d53f501e36175e900bf4986f44c7636d3b1b55df80804b92bb10d8da7798f6bb866722aa2354ad2880ab5c0f5c4633f198c137d2ca42b7c9
Currently the `fuzz.sh` script fails with "unbound variable" if called
without any arguments, this has gone undetected since we added `set
-euox pipefail` because in CI we always call it with an argument.
Use chatGPT to fix the bug.
Fix: #2924
1147e66896 Remove unnecessary braces (Jamil Lambert)
Pull request description:
Remove the unnecessary braces to eliminate the rust-analyzer warning,.
This is trivial, but it causes rust-analyzer to constantly flag a warning, and a simple change removes it.
I also added a comma at the end, again trivial but as someone who speaks Rust as a second language I have read this is what is normally done and I only mention it to get feedback on if that is correct or not.
ACKs for top commit:
Kixunil:
ACK 1147e66896
apoelstra:
ACK 1147e66896
Tree-SHA512: 454996e1ac8ac2a66ad855fe94eed63c9d0582f688a88ebfcda57a333129e7e50004d20c44819b97d2d56dc4bffca6580407262b39c37c4e878cff90092c76dc
Tobin C. Harding
merge-script
Martin Habovstiak
yancy
Ryan Breen
Jamil Lambert, PhD
Update Nightly Rustc Bot
Marko Bencun
Shing Him Ng