Re: Refactoring HMAC in the core code

On Sat, Dec 19, 2020 at 09:35:57AM +0900, Michael Paquier wrote:
> On Fri, Dec 18, 2020 at 10:48:00AM -0500, Bruce Momjian wrote:
> > Great. See my questions in the key manager thread about whether I
> > should use the init/update/final API or just keep the one-line version.
> > As far as when to commit this, I think the quiet time is actually better
> > because if you break something, it is less of a disruption while you fix
> > it.
>
> Please note that on a related thread that I have begun yesterday,
> Heikki has suggested some changes in the way we handle the opaque data
> used by each cryptohash implementation.
> https://www.postgresql.org/message-id/6ebe7f1f-bf37-2688-2ac1-a081d278367c@iki.fi
>
> As the design used on this thread for HMAC is similar to what I did
> for cryptohashes, it would be good to conclude first on the interface
> there, and then come back here so as a consistent design is used. As
> a whole, I don't think that there is any need to rush for this stuff.
> I would rather wait more and make sure that we agree on an interface
> people are happy enough with.

Others are waiting to continue working. I am not going to hold up a
patch over a one function, two-line API issue. I will deal with
whatever new API you choose, and mine will work fine using the OpenSSL
API until I do.

--
Bruce Momjian <bruce(at)momjian(dot)us> https://momjian.us
EnterpriseDB https://enterprisedb.com

The usefulness of a cup is in its emptiness, Bruce Lee