rangareddy opened a new pull request, #19433:
URL: https://github.com/apache/hudi/pull/19433
### Describe the issue this Pull Request addresses
Closes #16407 (HUDI-7455).
Seven bundles set
`<createDependencyReducedPom>false</createDependencyReducedPom>`, so the POM
published
to Maven Central still declares the artifacts that were shaded into the jar.
Consumers inherit those
artifacts as transitive dependencies even though the classes are already
inside the bundle — and are
relocated there, so the transitive copies are never the ones the bundle
itself uses.
What suggests this was unintended rather than deliberate: each of those
bundles *also* sets
`<dependencyReducedPomLocation>${project.build.directory}/dependency-reduced-pom.xml</dependencyReducedPomLocation>`
in the same `<configuration>` block, `dependency-reduced-pom.xml` is already
in `.gitignore`, and
`hudi-timeline-server-bundle` already sets the flag to `true` — as do the
nine bundles that simply don't
override shade's default.
### Summary and Changelog
Set `createDependencyReducedPom` to `true` for the seven bundles that had it
disabled:
`hudi-aws-bundle`, `hudi-hadoop-mr-bundle`, `hudi-datahub-sync-bundle`,
`hudi-presto-bundle`,
`hudi-azure-bundle`, `hudi-hive-sync-bundle`, `hudi-gcp-bundle`. No other
change; the reduced POM already
had a configured output location under `target/`.
Measured on `hudi-aws-bundle`, which declares six compile-scope dependencies:
- **before:** no reduced POM is generated at all, so all six are published
as transitive dependencies.
- **after:** `hudi-aws`, `hudi-hive-sync`, `httpclient`, `httpcore` are
dropped (all four are in the
bundle's `artifactSet`; the httpcomponents classes are relocated to
`org.apache.hudi.aws.org.apache.httpcomponents`), and `hudi-common` +
`parquet-avro` are kept, which is
correct — neither is bundled, so both are genuinely required at runtime.
All seven bundles were built (`mvn package -pl packaging/<bundle>
-DskipTests`) and each now emits a
reduced POM; `hudi-hive-sync-bundle` goes to one remaining dependency and
`hudi-hadoop-mr-bundle` to none.
`apache-rat:check` passes.
One scope note: the issue's headline example is `parquet-avro` published at
1.10.1 while bundled at 1.12.2
in 0.13.1. That specific mismatch no longer exists — `parquet-avro` is not
bundled any more, as the
reporter observed for 0.14.1 — so this PR intentionally does not remove it.
What it fixes is the
still-present leak of the artifacts that *are* bundled.
`hudi-io/pom.xml` also sets the flag to `false` and is deliberately left
alone: it is a library rather
than a shaded bundle, and whether its POM should be reduced is a separate
question.
### Impact
This changes the **published POM** of seven artifacts. A downstream build
that currently resolves, for
example, `httpclient` transitively through `hudi-aws-bundle` will need to
declare it directly. That is the
intended correction — a shaded bundle should not advertise dependencies it
has already absorbed — but it
is visible to consumers and worth a release note. Happy to stage it per
bundle instead of all seven at once
if you would prefer smaller steps.
No change to the jar contents, to any Hudi API, config, or table format.
### Risk Level
medium — it alters published dependency metadata for seven released
artifacts. Verified that all seven
build, that a reduced POM is produced for each, and (on `hudi-aws-bundle`)
exactly which dependencies are
removed versus retained, confirming shade only drops artifacts it actually
bundled. Relocation-vs-removal
was audited in detail for `hudi-aws-bundle`; for the other six I verified
the build and the reduced POM but
not every individual `<relocation>` entry.
### Documentation Update
none in the docs tree, but this is worth a line in the release notes for the
affected bundles, since
consumers may need to declare dependencies they previously received
transitively.
### Contributor's checklist
- [x] Read through [contributor's
guide](https://hudi.apache.org/contribute/how-to-contribute)
- [x] Enough context is provided in the sections above
- [x] Adequate tests were added if applicable
- [x] CI passes on my PR
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
