Bump k256

[neacsu]

cosmrs 0.8 release bumped k256, which makes cargo unable to select compatible versions for signature dependency crate when building cosmwasm-std and cosmrs together.

This PR also bumps k256dependency with what should be minimal effect on cosmwasm-crypto [1] to remove the said incompatibility.

[1] k256 CHANGELOG: https://github.com/RustCrypto/elliptic-curves/blob/master/k256/CHANGELOG.md

[ueco-jb]

Those changes depends on paragraph from digest CHANGELOG:
https://github.com/RustCrypto/traits/blob/master/digest/CHANGELOG.md#changed-1

Dirty traits are removed and instead block-level traits are introduced. Variable output traits reworked and now support both run and compile time selection of output size.

right?
I'd only ask you to update our CHANGELOG with entry about bumping k256 and digest to appropriate versions.

[neacsu]

Those changes depends on paragraph from digest CHANGELOG

Yes, that and a function rename.

I've updated the CHANGELOG as well.

[ueco-jb]

LGTM, thank you for contribution.

[shekohex]

@ueco-jb any plans on release a new version of this soon? like v1.0.1 for example?

[ueco-jb]

@shekohex I think right now we are quite close to releasing 1.1.0.
https://github.com/CosmWasm/cosmwasm/milestone/18
@webmaster128 or do you see anything worth adding to that milestone before release?

[webmaster128]

Looking at RustCrypto/signatures#409 (comment) and RustCrypto/utils#723 the core of this issue seems to be a MSRV problem, which causes ecdsa 0.13 and 0.14 being impossible to install in one project. This is very unfortunate but also not our bug. So I am hesitant to squeeze out a cosmwasm 1.0.1 patch release for that, especially given the non-trivial diff here (thank you by the way).

I think using latest master for your development and switching to 1.1.0 once it is out is a reasonable strategy.

[shekohex]

Sounds good, thank you @webmaster128 and @ueco-jb !