[FLINK-38268] Fix incorrect serialization of VARIANT type FLOAT

[JinkunLiu]

What is the purpose of the change

There is a small bug that appendFloat appends 8 bytes, which should be 4 instead. This may cause an ArrayIndexOutOfBoundsException when the buffer doesn't have enough capacity.

The page url is https://issues.apache.org/jira/browse/FLINK-38268

Brief change log

• FIX BinaryVariantInternalBuilder#appendFloat bug .

Verifying this change

Add a new unit test for FLOAT type. It would fail (throw an exception) without the fix.

Does this pull request potentially affect one of the following parts:

• Dependencies (does it add or upgrade a dependency): no
• The public API, i.e., is any changed class annotated with @Public(Evolving): no
• The serializers: yes
• The runtime per-record code paths (performance sensitive): no
• Anything that affects deployment or recovery: no
• The S3 file system connector: no

Documentation

• Does this pull request introduce a new feature? no

[flinkbot]

CI report:

• 84f8024 Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[davidradl]

Am I missing something? Long is 8 bytes and it is a writeLong. Could it be that checkCapacity(1 + 4); is incorrect and it should be checkCapacity(1 + 8); Line 286 is a write long with 8 bytes

[JinkunLiu]

The name of writeLong function might be misleading. The writeLong function does not always write 8 bytes. Its actual functionality is to write Float.floatToIntBits(f) into writeBuffer[writePos, writePos + numBytes). Similarly, you can refer to Line 198

[davidradl]

@JinkunLiu ok thanks for the clarification. It agree it is strange/misleading that the method is called writeLong, I would have expected the method to be called writeInt as it seems to have nothing to do with Longs and is writing an Int. Or writeBytes to be more generic.

[JinkunLiu]

hello @Sxnan, What do you think about the name of writeLong function.

[JinkunLiu]

hi @twalthr @Sxnan, I have fix this bug and add a new unitTest for appendFloat.

Could you help review those code? If I’ve made any mistakes, please feel free to point them out.