Re: SCRAM authentication, take three

On Wed, Feb 15, 2017 at 8:28 PM, Heikki Linnakangas <hlinnaka@iki.fi> wrote:
> On 02/07/2017 04:20 AM, Michael Paquier wrote:
>> --- a/src/backend/utils/errcodes.txt
>> +++ b/src/backend/utils/errcodes.txt
>> @@ -247,6 +247,7 @@ Section: Class 28 - Invalid Authorization
>> Specification
>>
>>  28000    E    ERRCODE_INVALID_AUTHORIZATION_SPECIFICATION
>> invalid_authorization_specification
>>  28P01    E    ERRCODE_INVALID_PASSWORD
>> invalid_password
>> +28P01    E    ERRCODE_INVALID_NONCE
>> invalid_nonce
>>
>
> Having two error codes with the same SQLSTATE is not cool, and tripped the
> assertion in PL/python. I removed the new error code, it was only used in
> one place, and ERRCODE_PROTOCOL_VIOLATIOn was more appropriate there anyway.
>
> Attached is a new set of patches, with that fixed. Thanks for the report
> Aleksander!

There is something that I think is still unwelcome in this patch: the
interface in pg_hba.conf. I mentioned that in the previous thread but
now if you want to match a user and a database with a scram password
you need to do that with the current set of patches:
local $dbname $user scram
That's not really portable as SCRAM is one protocol in the SASL
family, and even worse in our case we use SCRAM-SHA-256. I'd like to
change pg_hhba.conf to be as follows:
local $dbname $user sasl protocol=scram_sha_256
This is extensible for the future, and protocol is a mandatory option
that would have now just a single value: scram_sha_256. Heikki,
others, are you fine with that?
-- 
Michael