support for SCRAM-SHA-256 authentication

[codecov[bot]]

Codecov Report

Merging #58 (d7ea1e7) into master (96250d0) will increase coverage by 10.73%. The diff coverage is 75.14%.

@@             Coverage Diff             @@
##           master      #58       +/-   ##
===========================================
+ Coverage   53.63%   64.37%   +10.73%     
===========================================
  Files          13       14        +1     
  Lines        1251     1002      -249     
===========================================
- Hits          671      645       -26     
+ Misses        580      357      -223     

Impacted Files	Coverage Δ
src/pgo_connection.erl	45.09% <7.14%> (-2.53%)	:arrow_down:
src/pgo.erl	42.37% <20.00%> (-0.49%)	:arrow_down:
src/pgo_pool.erl	60.11% <70.00%> (+7.51%)	:arrow_up:
src/pgo_handler.erl	67.84% <71.55%> (+10.91%)	:arrow_up:
src/pgo_scram.erl	77.58% <77.58%> (ø)
src/pgo_type_server.erl	65.71% <78.57%> (-0.96%)	:arrow_down:
src/pgo_protocol.erl	55.96% <79.24%> (-1.89%)	:arrow_down:
src/pgo_sasl_prep_profile.erl	91.13% <91.13%> (ø)
... and 2 more

:mega: Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[lpil]

Is there a way to hide all these coverage annotations? It's pretty hard to read the code in places because of them

edit: found a way to do it per-file. Bit slow bit works

[tsloughter]

Oh yea, so tests pass on earlier version of postgres :(. Very confusing.

[tsloughter]

Something is screwy because now I also see:

Failure/Error: ?assertMatch(# { command := insert }, pgo : query ( "insert into hstore_tmp (id, labels) values ($1, $2)" , [ 1 , # { << "key-1" >> => << "value-1" >> , << "key-2" >> => << "value-2" >> } ] ))
  expected: = # { command := insert }
       got: {error,
                #{error => badarg_encoding,module => pg_types,
                  type_info =>
                      {type_info,16428,pg_array,[],default,
                          <<"_foo_ts_range">>,<<"array_send">>,
                          <<"array_recv">>,-1,<<"array_out">>,<<"array_in">>,
                          16429,
                          {type_info,16429,pg_record,[],default,
                              <<"foo_ts_range">>,<<"record_send">>,
                              <<"record_recv">>,-1,<<"record_out">>,
                              <<"record_in">>,0,undefined,0,
                              [23,3908],
                              [{type_info,23,pg_int4,[],default,<<"int4">>,
                                   <<"int4send">>,<<"int4recv">>,4,
                                   <<"int4out">>,<<"int4in">>,0,undefined,0,
                                   [],undefined},
                               {type_info,3908,pg_range,[],default,
                                   <<"tsrange">>,<<"range_send">>,
                                   <<"range_recv">>,-1,<<"range_out">>,
                                   <<"range_in">>,0,undefined,1114,[],
                                   undefined}]},
                          0,[],undefined},
                  value =>
                      #{<<"key-1">> => <<"value-1">>,
                        <<"key-2">> => <<"value-2">>}}}
      line: 358

And sometimes the timestamp tests pass with postgres 14. First guess is something is crashing but for some reason not showing up in the logs but results in configuration being lost on following tests that would explain the timestamp failures at least.

[benbro]

Not sure if it's relevant but fixing several silent or unhandled errors might help expose these kind of issues: https://github.com/erleans/pgo/issues/57 https://github.com/erleans/pgo/issues/42 https://github.com/erleans/pgo/issues/41

[tsloughter]

@benbro very good point.

[tsloughter]

@benbro this might have worked!

I'm still finishing up the commits because I'm not positive how best to handle particular failures but I should have a PR open today or tomorrow.

[tsloughter]

Found a bug related to the socket death not being handled right by the connection process too. Hopefully will have that PR ready sometime today.

[Neustradamus]

Linked to:

• https://github.com/scram-sasl/info/issues/1