Revert "signature: replace signature_derive with blanket impls (#1827)"

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it

[tarcieri]

This reverts commit bf47748.

Per #1831 this change breaks inference when there is a single explicit impl of the Signer trait.

It also wasn't possible to add corresponding blanket impls to the Async* traits, e.g. AsyncSigner for AsyncDigestSigner, because of the existing blanket impl of AsyncSigner for Signer which we definitely want to preserve.

As a general rule of thumb, blanket impls only make sense if they work 100% of the time, which doesn't seem to be happening here.

Closes#1831

[baloo]

Wait, which crates actually consumes the signature_derive again?

[tarcieri]

@baloo I think most of the crates avoid it to avoid pulling in the proc macro stack (in which case maybe it's not a great solution)

[baloo]

well where I'm trying to go is: maybe we can just pull the signature_derive altogether? I can't find a single consumer of it.

[tarcieri]

Sure

[baloo]

This is actually used in yubihsm.rs!

[tarcieri]

Yeah, that's probably the one thing that uses it