Closed baloo closed 1 month ago
Last set of CI errors are related to #799 I think.
cc @lumag
That last commit partially reverts #799 and #1010
I can't really find a way to keep the new_with_label
functions in RsaOaepParams
and RsaPssParams
I'm not sure I understand the purpose of this trait or why you're adding bounds for it, nearly all of which seem unused?
parameters of spki::AlgorithmIdentifier are serialized as Any.
...but ANY
represents... any type, so this is definitionally always true.
I guess your goal here is to try to constrain the parameters specifically to the der
crate's Any
and AnyRef
types....
...but that seems needlessly restrictive. The ASN.1 concept of ANY
allows for any type in that position. It's closer in concept to the original generic parameter.
Is there a specific motivation for this change?
Notably this precludes using e.g. PemReader
in a heapless no_std
environment to decode Params
to an owned, stack-allocated type, which seems like a regression.
The motivation was to compare the parameter to AnyRef::NULL
or Any::null()
(get access to is_null()
). But I guess I forgot the typesystem was already telling that to me :shrug:
Sorry for the noise and the distraction.
This newly defined
AnyLike
trait would help to make sure the parameters ofspki::AlgorithmIdentifier
are serialized as Any.