Skip to content

Releases: linkedin/avro-util

0.2.1

12 Jun 21:53
Compare
Choose a tag to compare

Release notes:

  • Improved fast serializer code generation logic for Enum type (#54)
  • Made fast serde generation deterministic (#57)
  • Added tests for primitive array serialization (#59)
  • New Primitive List APIs (#55)
  • Array-based implementation of the PrimitiveList APIs (#60)
  • Addressed JavaDoc build failures (#61)

0.1.19

27 May 18:00
10bef2f
Compare
Choose a tag to compare
add newBinaryDecoder which re-configures a BinaryDecoder in AvroCompa…

Bug fix in CompositeByteBuffer related to object reuse

13 May 20:32
4f0f61f
Compare
Choose a tag to compare

Bug fix in CompositeByteBuffer related to object reuse.

0.1.17: bug fix in fast-avro lib

06 May 22:28
84432ee
Compare
Choose a tag to compare
Fixed two bugs in FastDeserializerGenerator (#46)

1. fixed null exception when updating actual exceptions for already generated record method;
2. corrected value action symbols iteration order for nested Map;

Co-authored-by: Bingfeng Xia <bixia@linkedin.com>

0.1.16: test publishing artifact to Bintray

14 Apr 00:21
e49d36e
Compare
Choose a tag to compare
Fixed a bug in processUnion of FastSerializer (#44)

Previously, the code is checking schema type to decide the right index,
and this change was introduced to accommodate Avro-1.4, but this is
not appropriate since Union could contain multiple record types, which
will cause the writen index will always be the index of the first record.
The fix is to check the full name of the schema instead of type name.

0.1.15: Fast-avro - fixed a bug when processing union in Fast Serializer

13 Apr 23:18
e49d36e
Compare
Choose a tag to compare
Fixed a bug in processUnion of FastSerializer (#44)

Previously, the code is checking schema type to decide the right index,
and this change was introduced to accommodate Avro-1.4, but this is
not appropriate since Union could contain multiple record types, which
will cause the writen index will always be the index of the first record.
The fix is to check the full name of the schema instead of type name.

0.1.14: Bug fix in fast-avro

07 Apr 20:53
65ca605
Compare
Choose a tag to compare
Initialize ByteBuffer for PrimitiveFloatList List<Float> interface us…

0.1.13 : Get rid of jboss-logging-spi dependency

04 Apr 04:44
Compare
Choose a tag to compare
Implement WeakIdentityHashMap in avro-fastserde

0.1.12: Fast-avro improvements

02 Apr 18:31
a31b48f
Compare
Choose a tag to compare

There are two major improvements in the release about fast-avro:

  1. Utility functions to cache Fast DatumReader/DatumWriter to minimize the overhead of retrieving back Fast DatumReader/DatumWriter according to schemas;
  2. Lazy de-serialization by using ByteBuffer to improve the float array de-serialization by 40%;

0.1.11: remove avro-fastserde's declared dependency on helper from pom (#34)

24 Mar 03:59
7ae6344
Compare
Choose a tag to compare
Co-authored-by: Radai Rosenblatt <rrosenbl@rrosenbl-mn2.linkedin.biz>