secp256k1

secp256k1 copied to clipboard

group: ge(j) should have as invariant that the curve equation holds

3

I was surprised to see that this may be violated in `secp256k1_eckey_pubkey_parse`: https://github.com/bitcoin-core/secp256k1/blob/cc557575522c4cf11e5bcde1fea9637339cea21f/src/eckey_impl.h#L26-L31 I claim - `ge(j)` objects should always represent valid points on the curve. - This invariant should...

real-or-random

cmake: Use `SECP256K1_COVERAGE` option instead of `CMAKE_BUILD_TYPE=Coverage`

3

Providing `-DCMAKE_BUILD_TYPE=Coverage` does not work for multi-config generators. For example: ``` $ cmake -S . -B ../build -G "Ninja Multi-Config" $ cmake --build ../build -C Coverage ``` still builds the...

hebasto

Symbol visibility

11

In certain cases, we export too many symbols: - [ ] With `CFLAGS=-fvisibility=default`, there's a few exported _variables_ : `secp256k1_ecmult_gen_prec_table`, `secp256k1_pre_g`, `secp256k1_pre_g_128` - [x] on ARM, all the functions in...

real-or-random

fiat-crypto + CryptOpt tracking issue

29

[fiat-crypto](https://github.com/mit-plv/fiat-crypto) can generate verified field code for multiple targets, e.g., C and x86_64 asm. It has algorithm templates for our [mul](https://github.com/mit-plv/fiat-crypto/pull/1500) and [sqr](https://github.com/mit-plv/fiat-crypto/pull/1576) algorithm (under the name "Dettman") for secp256k1's...

real-or-random

cmake: Fix checking compiler flags like `-Wno-some-warning`

3

Some compilers ([GCC](https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html)) produce no diagnostic for `-Wno-some-warning` unless other diagnostics are being produced: ``` $ git diff diff --git a/CMakeLists.txt b/CMakeLists.txt index 240557f..f976824 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@...

hebasto

build: Disable some targets when external callbacks are enabled

1

Closes https://github.com/bitcoin-core/secp256k1/issues/1308.

hebasto

build: benchmarks and examples should be disabled when external callbacks are enabled

4

see https://github.com/bitcoin-core/secp256k1/pull/1306

real-or-random

More ARM Linux builds, fewer MacOS builds

4

Cirrus has native ARM64 Linux jobs, so we should use them (see also https://github.com/bitcoin-core/secp256k1/issues/707#issuecomment-902639436) And if we have that, and since MacOS jobs will run on ARM / M1 after...

real-or-random

release: Release tarballs? How to sign releases?

6

Some things that popped up during the 0.2.0 release: * Which tarball should we us as official release in the future? * github or * locally run `git archive` *...

real-or-random

Algorithm by Peter Dettman, with original comments: > Changes to _divsteps_59 (_30) that give maybe 4% speed improvement to const-time modinv on 64 bit. I see a larger gain on...

sipa