mlkem-native has been integrated into liboqs (see #653).
We should test in our CI that future changes in mlkem-native do not break that integration.
The approach should be similar to what is done for AWS-LC (#654).
Here are the steps needed for liboqs:
- clone the main branch of liboqs (if this is too flaky, we can fix to the last release version)
- Patch the copy_from_upstream.yml with the new commit.
- Run the copy_from_upstream.py (that has some python dependencies)
- Run the liboqs build as per README
- Once with
cmake -DOQS_DIST_BUILD=ON -DOQS_OPT_TARGET=auto -DCMAKE_BUILD_TYPE=Release (C backend)
- Once with
cmake -DOQS_DIST_BUILD=OFF -DOQS_OPT_TARGET=auto -DCMAKE_BUILD_TYPE=Release (native backend - should do this on both x86 and AArch64)
- Run tests
pytest -k ML-KEM
@bhess, does this make sense?
mlkem-native has been integrated into liboqs (see #653).
We should test in our CI that future changes in mlkem-native do not break that integration.
The approach should be similar to what is done for AWS-LC (#654).
Here are the steps needed for liboqs:
cmake -DOQS_DIST_BUILD=ON -DOQS_OPT_TARGET=auto -DCMAKE_BUILD_TYPE=Release(C backend)cmake -DOQS_DIST_BUILD=OFF -DOQS_OPT_TARGET=auto -DCMAKE_BUILD_TYPE=Release(native backend - should do this on both x86 and AArch64)pytest -k ML-KEM@bhess, does this make sense?