[
https://issues.apache.org/jira/browse/KAFKA-10389?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jason Gustafson updated KAFKA-10389:
------------------------------------
Summary: Fix zero copy tagged byte arrays (was: Fix zero copy tagged
arrays)
> Fix zero copy tagged byte arrays
> --------------------------------
>
> Key: KAFKA-10389
> URL: https://issues.apache.org/jira/browse/KAFKA-10389
> Project: Kafka
> Issue Type: Bug
> Reporter: Jason Gustafson
> Priority: Major
>
> We're missing a bunch of zero-copy checks in the message generator for tagged
> fields. As an example, try adding the following field to
> `SimpleExampleMessage.json`:
> {code}
> { "name": "taggedZeroCopyByteBuffer", "versions": "1+", "type": "bytes",
> "zeroCopy": true,
> "taggedVersions": "1+", "tag": 8, "ignorable": true },
> {code}
> The generated code has a few compilation errors because it assumes the byte
> array format. For example:
> {code}
> if (taggedZeroCopyByteBuffer.hasRemaining()) {
> _writable.writeUnsignedVarint(8);
>
> _writable.writeUnsignedVarint(this.taggedZeroCopyByteBuffer.length +
> ByteUtils.sizeOfUnsignedVarint(this.taggedZeroCopyByteBuffer.length + 1));
>
> _writable.writeUnsignedVarint(this.taggedZeroCopyByteBuffer.length + 1);
> _writable.writeByteArray(this.taggedZeroCopyByteBuffer);
> }
> {code}
> The `toStruct` and `fromStruct` methods also seem to be missing checks.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
