feat: Create keys from owned array values #781
Conversation
uncomputable
force-pushed
the
2025-02-array-params
branch
from
a0c356c to
9b66c2a
Compare
|
Deprecation of a method is a candidate for the changelog. |
Update TupleVisitor{32,33} to pass its owned byte array to the parsing
function instead of a mere byte slice. This gives us more flexibility
inside the parsing function.
Construct KeyPair directly from [u8; 32]. Deprecate KeyPair::from_seckey_slice and replace all of its calls with the new method.
uncomputable
force-pushed
the
2025-02-array-params
branch
from
9b66c2a to
e7eea32
Compare
|
Fixed formatting |
|
Do we have serde regression tests here? I'm pretty nervous that the first commit changes the serialization format. |
|
This PR leaves the serialization format unchanged. The changes in |
uncomputable
changed the title
|
I updated the other key constructors to take owned array values. I updated the PR title and added an entry to the changelog for a breaking change. @Kixunil I didn't see your approval until now; sorry. |
|
The lint failure is real, not sure about WASM. |
|
WASM we can ignore. We should probably just drop the CI job entirely until they fix my 18-month-old issue about wasm-pack not working at all in CI setups. |
|
Oh, crap. Side note: |
I updated the constructors to take owned array values of the form [u8; LEN] instead of taking array references of the form &[u8; LEN]. This makes the constructors more canonical. Already in this commit, I could remove a bunch of calls to `&` or `*`. Because this is a breaking change, I added an entry to the changelog.
uncomputable
force-pushed
the
2025-02-array-params
branch
from
8bd7f30 to
59cf7bd
Compare
|
WASM is already broken on In any case, I resolved the clippy lint. |
Thanks! I would appreciate that, though bear in mind that I will never test it locally until they fix |
Construct KeyPair directly from [u8; 32]. I can change the parameter to &[u8; 32] depending on the discussion in #780.