Schnorrsig API improvements

[real-or-random]

Should be merged before #995 if we want this.

I suspect the only change here which is debatable on a conceptual level is the renaming. I can drop this of course.

[real-or-random]

If you want to test SECP256K1_DEPRECATED, simply remove the "32" here again and look out for compiler warnings.

[michaelfolkson]

Concept ACK

I read the discussion on the renaming and it seems the only downsides are that the BIP needs to edited and a deprecation process of schnorrsign_sign is needed. Other projects (e.g. DLCs) seem to want to be able sign variable length messages and we'd like to facilitate other uses of this library if it doesn't present material downsides. I haven't tested yet but thanks for the guidance comment.

[real-or-random]

only downsides are that the BIP needs to edited

Yeah, in fact, this is not even a downside in the sense that the need to edit is caused by the renaming. There was consensus among the authors and contributors of the BIP already that the BIP should be amended to support varlen messages. It just hasn't been done so far.

[michaelfolkson]

Built successfully on MacOS Big Sur with --enable-dev-mode flag and ran tests.

If you want to test SECP256K1_DEPRECATED, simply remove the "32" here again and look out for compiler warnings.

Followed this instruction and indeed got the compiler warnings:

examples/schnorr.c:111:18: warning: 'secp256k1_schnorrsig_sign' is deprecated: Use secp256k1_schnorrsig_sign32 instead [-Wdeprecated-declarations]
    return_val = secp256k1_schnorrsig_sign(ctx, signature, msg_hash, &keypair, auxiliary_rand);
                 ^
./include/secp256k1_schnorrsig.h:136:3: note: 'secp256k1_schnorrsig_sign' has been explicitly marked deprecated here
  SECP256K1_DEPRECATED("Use secp256k1_schnorrsig_sign32 instead");
  ^
./include/secp256k1.h:179:55: note: expanded from macro 'SECP256K1_DEPRECATED'
#   define SECP256K1_DEPRECATED(_msg) __attribute__ ((__deprecated__(_msg)))
                                                      ^
1 warning generated.
  CCLD     schnorr_example

[michaelfolkson]

ACK 735d8be

Not familiar with this codebase but code looks fine to me. Mainly just renames.

[jonasnick]

concept ACK

For the record, I probably wouldn't rename sign at this point because it's not obviously better. I agree that it's more logical in terms of the spec, but for a user looking at the include file it's not immediately clear what the "32" suffix means and potentially causes confused searches for the "normal" signing function. But I'm ok with the change.

[jonasnick]

ACK b8f8b99

[sipa]

utACK b8f8b99