I'm making good progress on writing the Avro schema processor but I have a
question about recursive schemas.
For example, in the interop test schema we have...
{"name": "recordField", "type":
{"type": "record", "name": "Node",
"fields": [
{"name": "label", "type": "string"},
{"name": "children", "type": {"type": "array", "items":
"Node"}}]}}
There are only three data types that have names in Avro: records, enums and
fixed.
1. Can records, enums and fixed all be used as custom type specifications?
For example, in the interop schema there is an enum named "enumField" that
is a sibling to the "recordField" node. Does that mean that "recordField"
could have a field of type "enumField"?
2. Do we require that a custom type be defined 'before' it is used or can
people can use custom types prior to their definition?
I was up until 3am last night coding on Avro. I'm really looking forward to
showing you guys my work. I'll likely wait to push the changes until after
1.2 is out the door to minimize the release chaos.
-Matt