Segmentation fault when using float + union type

[luup2k]

Test Case 1 on Python 3.4.5:
Union data type: {"name": "age", "type": [ "float", "null" ] }

records = [{"name": "Dark Larry", "age": 1108311.1},]
type(records[0]['age']) #float class 
import quickavro

with quickavro.FileWriter("example.avro") as writer:
	writer.schema = {
		"type": "record",
		"name": "Person",
		"fields": [
			{"name": "name", "type": "string" },
			{"name": "age",  "type": [ "float", "null" ] }
		]
	}
	for record in records:
		writer.write_record(record)

Result: python interpreted is aborted and "Segmentation fault" is raised.

---

Test Case 2 on Python 3.4.5:
Union data type: {"name": "age", "type": "float" }

records = [{"name": "Dark Larry", "age": 1108311.1},]
type(records[0]['age']) #float class 
import quickavro

with quickavro.FileWriter("example.avro") as writer:
	writer.schema = {
		"type": "record",
		"name": "Person",
		"fields": [
			{"name": "name", "type": "string" },
			{"name": "age",  "type":  "float" }
		]
	}
	for record in records:
		writer.write_record(record)

Result: Avro file is generated.

[luup2k]

same behavior using Double instead float avro type

[ChrisRx]

Definitely let me know if this doesn't fix your issue or you should find some more. And thanks for filling out the issue!