Avro-2785: Update specs on how unions encoded

[anhldbk]

My PR addresses the following issue Avro-2785

[Fokko]

Can you resolve the conflicts @anhldbk ?

[anhldbk]

@Fokko What should I do? I agree with Roger's comment

[Fokko]

I can't merge the PR, there is a merge conflict in lang/py/avro/io.py. You should pull the latest master into your branch, and resolve the conflicts.

[anhldbk]

@Fokko PTAL :)

[Fokko]

@anhldbk I see a lot of unrelated changes in lang/py/avro/io.py, could you check? We like to minimize the changes, to avoid unrelated changes in version control.

[anhldbk]

@Fokko I've got troubles with my Python formatter. Gonna resolve it and tell you later

[anhldbk]

@Fokko actualluy my latest commit fixed the unrelated changes due to incompatibility in Python formatters. Please check

 λ git diff AVRO-2785 upstream/master -- lang/py/avro/io.py
diff --git a/lang/py/avro/io.py b/lang/py/avro/io.py
index e165db3a..b910ba56 100644
--- a/lang/py/avro/io.py
+++ b/lang/py/avro/io.py
@@ -846,7 +846,7 @@ class DatumReader(object):
def read_union(self, writers_schema, readers_schema, decoder):
"""- A union is encoded by first writing an int value indicating+ A union is encoded by first writing a long value indicating the zero-based position within the union of the schema of its value. The value is then encoded per the indicated schema within the union."""
@@ -1096,7 +1096,7 @@ class DatumWriter(object):
def write_union(self, writers_schema, datum, encoder):
"""- A union is encoded by first writing an int value indicating+ A union is encoded by first writing a long value indicating the zero-based position within the union of the schema of its value. The value is then encoded per the indicated schema within the union."""

[kojiromike]

Looks good to me. Thanks for contributing!

[Fokko]

Thanks @anhldbk