See https://mailarchive.ietf.org/arch/msg/spasm/b7rvHiKnA80oZ-VqQY7LIoFdsu8/. The point below is the that using a template like this with an embedded ASN.1 definition causes compilers (or at least the compiler I use) to emit far more code than if the structure were factored out of the template definition. I’ve not spent any time with composite-kem yet, but it looks like some of the comments I had relative to the composite-signatures draft would apply here too.
See https://mailarchive.ietf.org/arch/msg/spasm/b7rvHiKnA80oZ-VqQY7LIoFdsu8/. The point below is the that using a template like this with an embedded ASN.1 definition causes compilers (or at least the compiler I use) to emit far more code than if the structure were factored out of the template definition. I’ve not spent any time with composite-kem yet, but it looks like some of the comments I had relative to the composite-signatures draft would apply here too.