Re: [Dcrup] I-D draft-ietf-dcrup-dkim-crypto-06

[Alessandro Vesely <[e-mail]>]

On Sat 16/Sep/2017 14:02:13 +0200 James Cloos wrote:
>> It is the impression I've gotten from reading the various project lists.
> 
> Nikos recently posted that he skipped pure because curdle chose to avoid it.

Yes, that's the reply to the question Jeremy told us he was going to ask on
that list, on Sep 14:
https://lists.gnupg.org/pipermail/gnutls-help/2017-September/004387.html

My understanding is that curdle's decision is due to the fact that PureEdDSA
requires two passes over the input.  But then I didn't read the full details of
their decision (summer 2016):
https://www.ietf.org/mail-archive/web/curdle/current/msg00266.html

Dealing with generic functions, the GnuTLS manual does not expand much on why
gnutls_privkey_sign_hash [1] won't work.  It may be interesting to compare its
man page with that of nettle's ed25519_sha512_sign [2].  BTW, nettle is yet
another lib which doesn't seem to be going to implement the pre-hash variant
any time soon.

[1] GnuTLS
https://gnutls.org/manual/html_node/Abstract-key-API.html#gnutls_005fprivkey_005fsign_005fhash

[2] Nettle
https://www.lysator.liu.se/~nisse/nettle/nettle.html#EdDSA

hth
Ale