On Wed, Mar 8, 2017 at 10:39 PM, Robert Haas <robertmhaas@gmail.com> wrote:
> On Tue, Mar 7, 2017 at 10:01 PM, Michael Paquier
> <michael.paquier@gmail.com> wrote:
>> This way, we can be sure that two UTf-8 strings are considered as
>> equivalent in a SASL exchange, in our case we care about the password
>> string (we should care about the username as well). Without SASLprep,
>> our implementation of SCRAM may fail with other third-part tools if
>> passwords are not made only of ASCII characters. And that sucks.
>
> Agreed. I am not sure this quite rises to the level of a stop-ship
> issue; we could document the restriction.
I am not sure about that, what we have now on HEAD is still useful and
better than MD5.
> However, that's pretty ugly; it would be a whole lot better to get this fixed.
Agreed.
> I kinda hope Heikki is going to step up to the plate here, because I
> think he understands most of it already.
The second person who knows a good deal about NFKC is Ishii-san.
Attached is a rebased patch.
--
Michael
