acf9ac13e9
This is just a bad test. It constructs a preallocated context object by starting from a non-preallocated context object, in a way that can't be done by users (since it directly constructs a `Secp256k1` struct) and a way that is very difficult to unwind, because you wind up with two pointers to the same underlying context object, one a "preallocated" one and one a normal one. If you then drop the preallocated one, it will call `secp256k1_context_destroy`, forcing you to manually deallocate the other one. If you drop the normally-allocated one, you need to mem::forget the preallocated one to avoid calling `secp256k1_context_destroy` twice. The whole thing is pretty fragile. There is another unit test, `test_raw_ctx`, which gets into the same situation but using the public API, and demonstrates a few ways to get out of it. |
||
|---|---|---|
| .github/workflows | ||
| contrib | ||
| examples | ||
| githooks | ||
| no_std_test | ||
| secp256k1-sys | ||
| src | ||
| tests | ||
| .gitignore | ||
| CHANGELOG.md | ||
| Cargo-minimal.lock | ||
| Cargo-recent.lock | ||
| Cargo.toml | ||
| LICENSE | ||
| README.md | ||
| clippy.toml | ||
| rustfmt.toml | ||
README.md
Rust Secp256k1
rust-secp256k1 is a wrapper around libsecp256k1, a C
library implementing various cryptographic functions using the SECG curve
secp256k1.
This library:
- exposes type-safe Rust bindings for all
libsecp256k1functions - implements key generation
- implements deterministic nonce generation via RFC6979
- implements many unit tests, adding to those already present in
libsecp256k1 - makes no allocations (except in unit tests) for efficiency and use in freestanding implementations
Contributing
Contributions to this library are welcome. A few guidelines:
- Any breaking changes must have an accompanied entry in CHANGELOG.md
- No new dependencies, please.
- No crypto should be implemented in Rust, with the possible exception of hash functions. Cryptographic contributions should be directed upstream to libsecp256k1.
- This library should always compile with any combination of features on Rust 1.48.0.
Githooks
To assist devs in catching errors before running CI we provide some githooks. If you do not already have locally configured githooks you can use the ones in this repository by running, in the root directory of the repository:
git config --local core.hooksPath githooks/
Alternatively add symlinks in your .git/hooks directory to any of the githooks we provide.
Benchmarks
We use a custom Rust compiler configuration conditional to guard the bench mark code. To run the
bench marks use: RUSTFLAGS='--cfg=bench' cargo +nightly bench --features=recovery.
A note on non_secure_erase
This crate's secret types (SecretKey, Keypair, SharedSecret, Scalar, and DisplaySecret)
have a method called non_secure_erase that attempts to overwrite the contained secret. This
method is provided to assist other libraries in building secure secret erasure. However, this
library makes no guarantees about the security of using non_secure_erase. In particular,
the compiler doesn't have any concept of secrets and in most cases can arbitrarily move or copy
values anywhere it pleases. For more information, consult the zeroize
documentation.
Fuzzing
If you want to fuzz this library, or any library which depends on it, you will
probably want to disable the actual cryptography, since fuzzers are unable to
forge signatures and therefore won't test many interesting codepaths. To instead
use a trivially-broken but fuzzer-accessible signature scheme, compile with
--cfg=secp256k1_fuzz in your RUSTFLAGS variable.
Note that cargo hfuzz does not set this config flag automatically. In 0.27.0
and earlier versions, we used the --cfg=fuzzing which honggfuzz does set, but we
changed this because there was no way to override it.