You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
* AR uses the OID to map to named types for which handlers are registered
* JDBC can return the named types with getColumnTypeName()
* For normal (non-array, non-rage, etc.) this already works
* by adding handler based on the name as well as the OID,
getColumnTypeName() can be used for everything. OID based mapping
is still required for SchemaStatements.
* loading additional types needs to be adapted to handle OIDs as well
as type names, with or without namespace
* pgjdbc returns two additional type: 'serial' and 'bigserial'. They're
just int4 and int8 with auto-increment on top
* only trick thing is "fmod". luckily this is only used for "numeric"
and can be re-constructed by using getPrecision() and getScale()
* one test needs to be excluded because it checks for a message containing
an OID (number) instead of name (string) we get now
0 commit comments