[HADOOP-16822] Provide source artifacts for hadoop-client-api

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.3.1, 3.4.0, 3.2.3
• Fix Version/s: 3.3.1, 3.4.0, 3.2.3
• Component/s: build
• Labels: pull-request-available
• Target Version/s: 3.3.1, 3.4.0, 3.2.3
• Hadoop Flags: Reviewed
• Assignee: Karel Kolman
• Reporter: Karel Kolman
• Created: 21/Jan/20 10:30
• Updated: 11/Apr/24 14:30
• Resolved: 13/May/21 12:42

Description

Improvement request

The third-party libraries shading hadoop-client-api (& hadoop-client-runtime) artifacts are super useful.
 
Having uber source jar for hadoop-client-api (maybe even hadoop-client-runtime) would be great for downstream development & debugging purposes.

Are there any obstacles or objections against providing fat jar with all the hadoop client api as well ?

Dev links

• maven-shaded-plugin and its shadeSourcesContent attribute
• https://maven.apache.org/plugins/maven-shade-plugin/shade-mojo.html#shadeSourcesContent

Update April 2024: this has been reverted.

It turns out that it complicates debugging. If you want the source when debugging, the best way is just to check out the hadoop release you are working with and point your IDE at it.

Issue Links

• relates to: HADOOP-11804 Shaded Hadoop client artifacts and minicluster
• relates to: HADOOP-13989 Remove erroneous source jar option from hadoop-client shade configuration
• links to: GitHub Pull Request #6458
• links to: GitHub Pull Request #6719