Update to swift-crypto 2.4.1

[yim-lee]

Motivation:
swift-crypto 2.4.1 includes updates in BoringSSL that cause Unsafe*Pointer<EVP_PKEY> become OpaquePointer and without this code adjustment SwiftPM fails to build with error: cannot find type 'EVP_PKEY' in scope.

Modifications:

• Update dependency from 2.4.0 to 2.4.1
• Replace Unsafe*Pointer<EVP_PKEY> with OpaquePointer.

[yim-lee]

This requires changes to update-checkout.

[yim-lee]

swiftlang/swift#65157
@swift-ci please smoke test

[MaxDesiatov]

Ready to merge? This would unblock a couple of other pull requests.

[yim-lee]

Ready to merge? This would unblock a couple of other pull requests.

Need @shahmishal to merge the update-checkout changes, otherwise this patch would break other CI.

[neonichu]

BTW, what will happen in swift-crypto 2.4.2? Is this considered a bug that'll get fixed (which implies we should use exact still) or will it stay this way?

[yim-lee]

Is this considered a bug that'll get fixed

No, it's not a bug. It is a part of BoringSSL updates that this version of swift-crypto took in, and will stay.

swift-crypto updates its vendored copy of BoringSSL once a month or so. Usually the changes are fairly minor and non-breaking, but sometimes (like this one) the updates are big and cause problems. The breaking change from swift-crypto 1.x to 2.x was similar to what happened here.

No one is supposed to depend on CCryptoBoringSSL really, but we wrote PackageCollectionsSigning using it because that was the best option at the time. Now that we have swift-certificates, the long term solution is to rewrite PackageCollectionsSigning with it and remove the dependency on CCryptoBoringSSL. I am working on that for main.

In the meantime and for 5.9, SwiftPM could as a safety measure keep using exact version (2.4.1). @neonichu @tomerd @MaxDesiatov Do we want to do that? If so I can update this and the PR for 5.9 to use exact.

[neonichu]

Oh I see, thanks for the explanation.

I don't have a strong opinion, this seems rare enough that we can live with it. Just wanted to make sure we don't have a guaranteed breakage in the near future.

For 5.9 specifically it might make sense to pin if we are somewhat certain we won't need to update it again.

[MaxDesiatov]

Committed conflicting PR just to unblock main in the meantime, but I resolved the conflict here, I hope you don't mind.

[MaxDesiatov]

swiftlang/swift#65157
@swift-ci please smoke test

[MaxDesiatov]

swiftlang/swift#65157
@swift-ci test windows

[MaxDesiatov]

@swift-ci test windows

[MaxDesiatov]

swiftlang/swift#65157
@swift-ci test windows

[MaxDesiatov]

swiftlang/swift#65157
@swift-ci test windows

[MaxDesiatov]

@swift-ci test windows

[MaxDesiatov]

Windows regression introduced in swiftlang/swift#65085

[shahmishal]

@swift-ci test windows

[yim-lee]

@swift-ci please test Windows platform

[yim-lee]

swiftlang/swift#65157
@swift-ci please test Windows platform

[yim-lee]

swiftlang/swift#65157
@swift-ci please test Windows platform

[MaxDesiatov]

@swift-ci test windows