0046bb8ad8 Fix usage of cfg(rust_1_53) (Tobin C. Harding)
c3450f3913 Remove stale usage of doc(cfg) (Tobin C. Harding)
Pull request description:
These build cfg options are not features, fix broken usage. And remove stale docsrs attribute while we are at it. Bad rust-bitcoin devs.
Found while reviewing #1870
ACKs for top commit:
apoelstra:
ACK 0046bb8ad8
Kixunil:
ACK 0046bb8ad8
Tree-SHA512: 7053affef6654ff203c93590bf081e165f019feb040aa8c55259ffe4e15eaf0e7522f6e5a4f6f62e8f578420b0313f4b7b17c46b741b7fcfd05750e5c5976589
3af9258025 embedded: Document how to clean up linker flags (Tobin C. Harding)
Pull request description:
Our embedded crate includes instructions to source a shell script that sets the `RUSTFLAGS` env var. Having the env var set like this in ones environment breaks linkage when trying to do "regular" builds.
Document how to clean up.
ACKs for top commit:
apoelstra:
ACK 3af9258025
Kixunil:
ACK 3af9258025
Tree-SHA512: 72758fba4dede873da299f01d75fd64b549fe21f954c2720ba3b7fc3c29fd4ed28fd0a749bbe987a7de1551aa32192253dd0033a18c877e877c9960343a5b07c
Our embedded crate includes instructions to source a shell script that
sets the `RUSTFLAGS` env var. Having the env var set like this in ones
environment breaks linkage when trying to do "regular" builds.
Document how to clean up.
013dffa65d tests: Use script hash methods (Tobin C. Harding)
Pull request description:
The `ScriptBuf` type can be serialized using it's `to_bytes` function. Do not use the `psbt::Serialize` trait to do so in test code.
No logic changes, since the impl of `psbt::Serialize` for `ScriptBuf` just calls `to_bytes`.
ACKs for top commit:
Kixunil:
ACK 013dffa65d
apoelstra:
ACK 013dffa65d
Tree-SHA512: 08959e065f1528f2ca69c12d5e34bceea3ddd17eefcee45094f071b3fa7357dbf289f6fe54d18fea02eecd1d0a7cd04598bbf014a5f10676387dbe27bb490395
b03c24db8c Add a checked version of weight mul fee_rate (yancy)
Pull request description:
Add a checked version of fee_rate * weight. While I like the trait version of just being able to multiply `feerate * weight`, it's not really very useful imo since a large input feerate could cause an overflow. Instead of changing the trait in https://github.com/rust-bitcoin/rust-bitcoin/pull/1849 (not idiomatic enough I guess) I added a `checked_weight_mul` method to `FeeRate`.
ACKs for top commit:
apoelstra:
ACK b03c24db8c
Kixunil:
ACK b03c24db8c
Tree-SHA512: 231ade94291becadcea9ea2a40a5daf96b77f01a29cca2494d7bbe4f7de5b412fa8fc816ea249268569f5378410185d9349fd687533bf3a422a752997e107a2b
We have methods to convert a script to a `WScritpHash` and `ScriptHash`,
no need to do this manually, let alone use the `psbt::Serialize` trait
to do so.
During the last round of releases (bitcoin 0.30, hashes 0.12) we removed
the `FromHex` implementation from all types except vecs and arrays. We
added `FromStr` impls for types that roundtrip with `Display`.
We never added a changelog mention to either `bitcoin` or `hashes`, lets
retroactively add an entry.
Fix: #1747
These constants had an error that they had `script_size` set to 0 which
was incorrect because it's not length of the script but serialized size.
Rather than just bumping the value this uses the `from_slice` method
which is less error-prone.
This also deletes a useless test of the constants.
Closes#1834
8835d5d2f1 make bip21 schema lowercase (Riccardo Casatta)
Pull request description:
The spec RFC3986 specifies the scheme is case insensitive and we were uppercasing to optimize QR code representation.
Unfortunately, common platform such as Android seems to fail to recognize uppercase schema, so for compatibility reason we use lowercase.
close#1843
ACKs for top commit:
Kixunil:
ACK 8835d5d2f1
apoelstra:
ACK 8835d5d2f1
Tree-SHA512: 02d228b52fe4df20edb71ba8e2ab8a2bae4b912252e30a3150ee3af74e65a6e91b165c9579273b57e894366c9792a8312ea973723cd8c5d98037aaba80d7cf07
6c6a89b1d1 Add sub-sat fractions parsing regression test (Martin Habovstiak)
f1a3dc6719 Allow parsing sub-sat denoms with decimal points (Martin Habovstiak)
b3d9a267ea Add a few more amount parsing tests (Martin Habovstiak)
Pull request description:
Numbers with only zeros after decimal points are valid if they are also
multiples of `10^precision` (e.g. 1000 for msats). These were
artificially disallowed as "too precise" which was at least misleading.
This change allows parsing such numbers.
And yes, I know this is not perfectly efficient (unless the compiler figures out some magic opts) but so isn't the rest of the code. TBH this parsing code drives me crazy and I'd love to rewrite it to be more efficient and readable.
ACKs for top commit:
apoelstra:
ACK 6c6a89b1d1
Tree-SHA512: 03cf4b416f2eac25e0aac57ef964ed06fa36c7fe8244bdcf97852cc58e1613b1ec6132379b834da58ad3240fdd61508a384202f63aa9ffa335c18cd7b2b724d3
The spec RFC3986 specifies the scheme is case insensitive and we were uppercasing
to optimize QR code representation.
Unfortunately, common platform such as Android seems to fail to recognize
uppercase schema, so for compatibility reason we use lowercase.
75b3f19b96 Move and rename TxOut default trait to a const called NULL (yancy)
Pull request description:
Create an associated constant `const TxOut::NULL` for consensus signing code and remove the default trait. Note I tried to deprecate the `default()` fn instead of just removing it but it doesn't seem to be possible. Also because `TxOut::NULL` is `const`, `ScriptBuf::new()` needed to be changed to `const fn`.
ACKs for top commit:
apoelstra:
ACK 75b3f19b96
Kixunil:
ACK 75b3f19b96
Tree-SHA512: ff61a2b1641a1ba32f183c27205af2d868dbc2eb47cf758c3d8315329d2c23e0b8a82ea0ab59d1de9add0d238f927165e2e4df014aab1ef066d74d4feda0700b
995c797e0d feat: generate PrivateKey (kshitjj)
Pull request description:
added a function to generate a private key
Resolves: #1823
ACKs for top commit:
apoelstra:
ACK 995c797e0d
tcharding:
ACK 995c797e0d
Tree-SHA512: 29ba54be8cb777e71a4683835686cbf2978b23736f629d7bbff468074235fece261ca170c23f358d1bd878987566d09e4488c3f1a106c59a5c8bdf52b98abffe
Numbers with only zeros after decimal points are valid if they are also
multiples of `10^precision` (e.g. 1000 for msats). These were
artificially disallowed as "too precise" which was at least misleading.
This change allows parsing such numbers.
8e6f953aa7 Expose valid (min, max) difficulty transition thresholds (Wilmer Paulino)
Pull request description:
Once `U256` was made private, we lost the ability to check whether a valid difficulty transition was made in the chain, since `Target` doesn't expose any operations. We only choose to expose `Shl<u32>` and `Shr<u32>` such that we can compute the min and max target thresholds allowed for a difficulty transition.
This is something we realized was missing after bumping to `rust-bitcoin v0.30.0` in `rust-lightning`, specifically for our `lightning-block-sync` crate. It may also be worth having a helper in `rust-bitcoin` that checks a header properly builds upon the previous, but that can be left for future work.
ACKs for top commit:
Kixunil:
ACK 8e6f953aa7
sanket1729:
ACK 8e6f953aa7 . Sorry, was confused by some details.
apoelstra:
ACK 8e6f953aa7
Tree-SHA512: 740dd64089426463dc6a19726d5a562276bd0966e0e31af8e1e67b28d18945644ac0e50be3cf0cc7fa604acc3d2c5b912a77a7caa69d8cff85f70fd57e5328c5
dff757d7db Comment predict_weight (yancy)
Pull request description:
I've been reading over the `predict_weight` function since it is one of the biggest challenges for coin-selection. IE choosing inputs and constructing an optimal selection strategy requires predicting the weight to get the best selection. It's great this work has been done but there are some things I don't understand well enough to comment.
1) why are we looking at the size of VarInt struct here
> let script_size = script_len + VarInt(script_len as u64).len()
2) [predict_weight_internal](36500b4451/bitcoin/src/blockdata/transaction.rs (L1245)) has a bunch of magic numbers. I'd like to be able to comment this as well but I don't fully understand that function.
Also, `Transaction.rs` is a big file and it seems like all of the prediction stuff could be moved to a separate module or maybe a separate crate?
ACKs for top commit:
tcharding:
ACK dff757d7db
Kixunil:
ACK dff757d7db
Tree-SHA512: 8ffa16d500075d691528ce1819b9352a148af431889bebbd7cddcf470bd4e3048ec53a5e778bc3659e33d8c25b68422a93dac1d46b9489ff56f41d88d7f05433
d57ec019d5 Use Amount type for TxOut value field (yancy)
Pull request description:
Propose using `Amount` type for the `TxOut` `value` field. I only implemented `Decodable ` and `Encodable` enough to compile but this needs to completed obviously if using `Amount` seems like a good idea.
ACKs for top commit:
tcharding:
ACK d57ec019d5
apoelstra:
ACK d57ec019d5
Tree-SHA512: df3fd55294d5f9392ca90bb920be8fbb9d7d285d97669412e07d5a099f70f81fd73e7e259679de9c8ce5c6c855e64f62213700f0fb7db415e0c706c509485377
ed6421c939 address: Add generic serde::Serialize for Address (Steven Roose)
814b9917da address: Add Sync, Send, Sized and UnPin marker traits on NetworkValidation (Steven Roose)
Pull request description:
With the new rewrite of Address, `serde::Serialize` is only implemented on `Address<bitcoin::address::NetworkChecked>` and `Address<bitcoin::address::NetworkUnchecked>`. But the compiler has no way of knowing that that are all the possible versions of `Address`, so the generic `Address<impl bitcoin::address::NetworkValidation>` doesn't implement `serde::Serialize`.
ACKs for top commit:
Kixunil:
ACK ed6421c939
tcharding:
ACK ed6421c939
Tree-SHA512: 65e43dff244c94fe08ccb2d985781a2687a1e2db186960a35d4ae89f3b31c5af66892630a3ebaac9cecdc83638487425afa17374869d278648b348869e0ba091
Once `U256` was made private, we lost the ability to check whether a
valid difficulty transition was made in the chain, since `Target`
no longer exposes any arithmetic operations.
6cab7beba3 Deprecate min/max_value methods (Tobin C. Harding)
5fbbd483ea Use MIN/MAX consts instead of min/max_value (Tobin C. Harding)
3885f4d430 Add MIN/MAX consts to amounts (Tobin C. Harding)
Pull request description:
The new MSRV (1.48.0) uses associated consts MAX/MIN instead of functions, we had functions to be compliant with the old MSRV.
~Remove all methods `min_value` and `max_value` including calls to these methods on stdlib types.~
PR is now split into three patches:
- patch 1: Add missing associated consts MIN/MAX as needed
- patch 2: Use consts instead of method calls
- patch 3: Deprecate methods `min_value` and `max_value`
ACKs for top commit:
sanket1729:
ACK 6cab7beba3
apoelstra:
ACK 6cab7beba3
Kixunil:
ACK 6cab7beba3
Tree-SHA512: 60949d1bb971e0dfbab7f573b4447f889b5fa1a5f1c9ac9325a2970fe17a19ccc93418dba57f07bed7e13864b130de48b6b3741d1d80266c6144237dd4565ff7
c4c64c0dc5 Test with minimal dependency versions (Martin Habovstiak)
d5655d503a Bump core2 dependency from 0.3.0 -> 0.3.2 (Tobin C. Harding)
Pull request description:
This is work originally done by Kixunil in #1272, I picked it up to help out. The only changes I made were rebasingg, updating the recent lock file, adding `--locked` to hashes contrib file, and adding a co-developed-by tag for accountability.
It could happen that we unknowingly depend on a new version of a crate without updating `Cargo.toml`. This could cause resolution issues for downstream users. It's also unclear for outsiders to know with which dependencies did we test the crate.
This change commits two lock files: `minimal` and `recent`. `minimal` contains minimal dependency versions, while `recent` contains dependency versions at the time of making the change.
Further, this adds CI jobs to test with both lock files, CI job for `internals` crate, removes old `serde` pinning and prints a warning if `recent` is no longer up to date. (We may have to override it somehow if any crate breaks MSRV.)
The documentation is also updated accordingly.
Closes#1230
ACKs for top commit:
apoelstra:
ACK c4c64c0dc5
Kixunil:
ACK c4c64c0dc5
Tree-SHA512: 7d386e96ab747f6a6bafeea828ac65bd8bb11975eaa3408acecac369cd2f235f6e9d4c57202be18a3dc2eeb2a2df532d73e4d35cd1f3fbf092eb6414c55b1524
Our previous MSRV did not support MIN/MAX associated consts so we had
methods min/max_value. Now that our MSRV is Rust 1.48.0 we can use the
consts.
Deprecate min/max_value methods in favor of MIN/MAX associated conts.
We currently use the functions `min_value` and `max_value` because the
consts were not available in Rust 1.41.1, however we recently bumped the
MSRV so we can use the consts now.
It could happen that we unknowingly depend on a new version of a crate
without updating `Cargo.toml`. This could cause resolution issues for
downstream users. It's also unclear for outsiders to see which
dependencies we tested the crate with.
This change commits two lock files: `minimal` and `recent`. `minimal`
contains minimal depdendency versions, while `recent` contains
dependency versions at the time of making the change.
Further, this adds CI jobs to test with both lock files, CI job for
`internals` crate, removes old `serde` pinning and prints a warning if
`recent` is no longer up to date. (We may have to override it somehow if
any crate breaks MSRV.)
The documentation is also updated accordingly.
Co-developed-by: Tobin C. Harding <me@tobin.cc>
Closes#1230
1c3bbd4bf2 internals: Remove attribution from all files (Tobin C. Harding)
99673ab5c4 hashes: Introduce SPDX license identifiers (Tobin C. Harding)
984fe69448 bitcoin: Remove attribution from all files (Tobin C. Harding)
Pull request description:
Please note, whether or not we need a per-file license comment is out of scope for this PR. This PR leaves us with the most simple per-file solution possible and leaves the merit of per-file license comment to be discussed on another day.
Simplify the per-file license stuff by doing:
- Remove the attribution line from each file.
Currently we have a mishmash of attribution lines accompanying the SPDX identifier. These lines are basically meaningless because:
- The date is often wrong
- The original author attributed is not the only contributor to a file
- The term "rust bitcoin developers" is basically just noise
- Introduce SPDX license identifiers into `hashes` and remove attribution line (ie, make `hashes` uniform with `bitcoin`)
Required before merge please:
- [x] ack from apoelstra because as the library original author many of the changes in this PR remove his name
- [x] ack from Kixunil because he had some concerns in the issue descussion
Fix: #1816
ACKs for top commit:
Kixunil:
ACK 1c3bbd4bf2
sanket1729:
ACK 1c3bbd4bf2
apoelstra:
ACK 1c3bbd4bf2
Tree-SHA512: c5ac05c5eb23b3b6a760f707c344b22f5871a4dedee4990b1840f57e4cee1d38560ff4507c354bbf29bc8ff05a179d95d7e100fcf19bd93c5362344a352c7b5a
Currently we have a mishmash of attribution lines accompanying the SPDX
identifier. These lines are basically meaningless because:
- The date is often wrong
- The original author attributed is not the only contributor to a file
- The term "rust bitcoin developers" is basically just noise
Just remove all the attribution lines and be done with it. While we are
at it add an SPDX line to the few files missing it, whether this license
nonsense is even needed is left as an argument for another day.
dd4ad9444e Hardcode expected weight in txin_txout_weight_tests (Peter Todd)
Pull request description:
Rational: the expected weight is fixed so this both ensures we don't accidentally change it somehow, and makes it easier to re-use these test cases in other codebases (eg python-bitcoinlib).
ACKs for top commit:
apoelstra:
ACK dd4ad9444e
tcharding:
ACK dd4ad9444e
Kixunil:
ACK dd4ad9444e
Tree-SHA512: 4769a4bb8695f4f4c95e258bb5f06a232090b14c3d9159d6d5de2d09d7fc934a1b920b90cc09677a88fc0cf37ac21ed27794692dff2c73df4252c9551dc10fc2
2860aae1a5 fuzz: don't fuzz hashes against RustCrypto (Andrew Poelstra)
6467728202 fuzz: disable tests unless 'cfg(fuzzing)' is passed; update README for reproducing failures (Andrew Poelstra)
6e2ee5be66 fuzz: run 'cargo fmt' on all the fuzz targets (Andrew Poelstra)
9cfc0fcd81 fuzz: add contrib/test.sh so we at least 'cargo test' it in CI (Andrew Poelstra)
933ecb19e1 fuzz: fix warnings, clippy lints, 1.48.0 failures (Andrew Poelstra)
fd88e48696 fuzz: remove AFL support (Andrew Poelstra)
ab467cb091 fuzz: make hongfuzz fuzzing the default feature (Andrew Poelstra)
6f754df231 fuzz: add fuzzing README (Andrew Poelstra)
f093765efe fix fuzz.sh and cycle.sh to use generated lists of targets (Andrew Poelstra)
6534f22362 fuzz: auto-generate CI and Cargo.toml files (Andrew Poelstra)
8021034d86 rename travis-fuzz.sh to fuzz.sh; partially patch CI (Andrew Poelstra)
0be75f7edc move hashes/fuzz into main fuzz/ directory (Andrew Poelstra)
5a891dec2d move bitcoin fuzz targets into bitcoin/ subdirectory (Andrew Poelstra)
e3111c748b move bitcoin/fuzz into repo root; add to workspace (Andrew Poelstra)
Pull request description:
Several big changes here:
* Moves fuzzing to its own workspace with a `contrib/test.sh` etc so that CI will check that it compiles
* FIx all warnings, clippy lints, MSRV problems, etc.; mostly move to Rust 2018
* Merge `hashes/` fuzztests into workspace
* Rewrite all scripts; add file that auto-generates CI fuzz job and Cargo.toml so we don't have to manually keep these in sync
* Remove bitrotted and partial AFL support.
Supercedes #1422
I suspect the hashes fuzztests will actually fail since we haven't touched them in so long. Will address that if CI fails here.
ACKs for top commit:
sanket1729:
ACK 2860aae1a5
tcharding:
ACK 2860aae1a5
Tree-SHA512: b1aa3d6fac75fee51966f1d3f3245784e331bdea2a3fa7d6609bc4196c34f81acb7701faf8f269c3741568ea100438f24a2f06e75c8d01cb84c8b22d7886f1dd
Rational: the expected weight is fixed so this both ensures we don't
accidentally change it somehow, and makes it easier to re-use these test
cases in other codebases (eg python-bitcoinlib).
Seems we no longer need an explicit error handler, remove it.
I did not grok the reason (long thread link below) but just removed it
and checked that the embedded crates still ran correctly.
https://github.com/rust-lang/rust/issues/51540)
Andrew Poelstra
Tobin C. Harding
yancy
Martin Habovstiak
Riccardo Casatta
kshitjj
Wilmer Paulino
Peter Todd