Re: Improve error handling of HMAC computations and SCRAM

On Tue, Jan 11, 2022 at 11:08:59AM +0300, Sergey Shinderuk wrote:
> Yeah, that's better.  I thought "providing errors about an error" was a
> typo, but now I see the same comment was committed in b69aba745.  Is it just
> me? :)

It is not only you :)  I have applied a fix to fix the comments on
HEAD and REL_14_STABLE.

Attached is a rebased patch for the HMAC portions, with a couple of
fixes I noticed while going through this stuff again (mostly around
SASLprep and pg_fe_scram_build_secret), and a fix for a conflict
coming from 9cb5518.  psql's \password is wrong to assume that the
only error that can happen for scran-sha-256 is an OOM, but we'll get
there.
--
Michael