Add pg_get_result_enc_idx

[jeremyevans]

This can be used by sequel_pg to get the encoding index, as discussed in jeremyevans/sequel_pg#34.

I've prepared an update to sequel_pg to use this function:

$ ruby24 -I /data/code/external/ruby-pg/lib bin/sequel postgres:///?user=sequel_test -c 'p $LOADED_FEATURES.grep(/pg.*.so\z/); p DB[:s].get(:s)'
["/data/code/external/ruby-pg/lib/pg_ext.so", "/usr/local/lib/ruby/gems/2.4/gems/sequel_pg-1.12.2/lib/sequel_pg.so"]
"\xE1\x80\x80"
$ ruby24 -I /data/code/sequel_pg/lib -I /data/code/external/ruby-pg/lib bin/sequel postgres:///?user=sequel_test -c 'p $LOADED_FEATURES.grep(/pg.*.so\z/); p DB[:s].get(:s)'
["/data/code/external/ruby-pg/lib/pg_ext.so", "/data/code/sequel_pg/lib/sequel_pg.so"]
"\u1000"

I would appreciate it if pg 1.2.1 could be released with this change.

[ged]

LGTM. I'll push out a release ASAP, as my company is blocked on this too. ;)