[FLINK-33725][core] Return false from MathUtils.isPowerOf2 for non-positive values

[vasiliy-mikhailov]

What is the purpose of the change

MathUtils.isPowerOf2(long) returns true for 0 and for Long.MIN_VALUE,
neither of which is a power of two. 0 & (0 - 1) == 0 and
Long.MIN_VALUE & (Long.MIN_VALUE - 1) == 0, so the existing bit trick alone
misclassifies non-positive inputs. This fixes
FLINK-33725.

Brief change log

• Guard isPowerOf2 with value > 0 so non-positive values return false.
• Extend MathUtilTest#testPowerOfTwo with the 0, -1 and Long.MIN_VALUE cases.

Verifying this change

Added assertions to MathUtilTest#testPowerOfTwo for 0, -1 and Long.MIN_VALUE. Running just that test on the current master (before the one-line change) fails, and passes after it:

Before the fix (on master):

$ mvn test -Dtest=MathUtilTest#testPowerOfTwo -pl flink-core
[INFO] Running org.apache.flink.util.MathUtilTest
[ERROR] Tests run: 1, Failures: 1, Errors: 0, Skipped: 0 <<< FAILURE! -- in org.apache.flink.util.MathUtilTest
[ERROR] org.apache.flink.util.MathUtilTest.testPowerOfTwo <<< FAILURE!
Expecting value to be false but was true
	at org.apache.flink.util.MathUtilTest.testPowerOfTwo(MathUtilTest.java:126)
[INFO] BUILD FAILURE

After the fix:

$ mvn test -Dtest=MathUtilTest#testPowerOfTwo -pl flink-core
[INFO] Running org.apache.flink.util.MathUtilTest
[INFO] Tests run: 1, Failures: 0, Errors: 0, Skipped: 0 -- in org.apache.flink.util.MathUtilTest
[INFO] BUILD SUCCESS

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: no
• The runtime per-record code paths (performance sensitive): no
• Anything that affects deployment or recovery: JobManager (and its components), Change to the configuration of TaskManager, etc.: no
• The S3 file system connector: no

Documentation

• Does this pull request introduce a new feature? no
• If yes, how is the feature documented? not applicable

Was generative AI tooling used to co-author this PR?

• Yes (Claude Code, Claude Opus 4.8)

Generated-by: Claude Code (Claude Opus 4.8)

[flinkbot]

CI report:

• f69544b Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

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

[spuru9]

Changes look good, Can you update the PR description as per guidlines.

[vasiliy-mikhailov]

Thanks for taking a look. I've updated the description to follow the PR template (filled in the full "potentially affect" checklist and the verifying section).

[spuru9]

Thanks for taking a look. I've updated the description to follow the PR template (filled in the full "potentially affect" checklist and the verifying section).

Can you add the AI declaration too?

[vasiliy-mikhailov]

Added an AI assistance disclosure to the description. For transparency: an AI pipeline was used to scan a large amount of code, surface suspected bugs, reproduce a subset with failing unit tests and generate fixes, and prepare PRs; each one — including this — was reviewed and verified by a human (fail-before / pass-after test run plus code review) before opening.

[spuru9]

Can you refer #28463 (comment)

[vasiliy-mikhailov]

Thanks @spuru9 — good call, I've switched to the standard template field from #28463: checked "Was generative AI tooling used to co-author this PR?" and added the Generated-by: trailer, replacing my earlier free-form note.