Per-jar META-INF/LICENSE should describe each jar's actual contents

[bobbai00]

Task Summary

Each Texera jar currently ships with a copy of the repo-root LICENSE file in its META-INF/LICENSE. This is problematic because the LICENSE is not customized for what each jar actually contains.

Per ASF policy, a jar's META-INF/LICENSE must describe what is in that specific jar.

• workflow-operator jars should have LICENSE containing Apache 2.0 plus the MIT attribution and license text. , because it has 3rd-party code from mbknor-jackson-jsonschema (using MIT license)
• Other jars don't have any 3rd party code. So they should have LICENSE containing only the Apache 2.0 license text.

Here is a table summarize what jars Texera will produce, 3rd party code inside it, and what content should be contained in the META-INF/LICENSE:

Jar	Vendored third-party code	META-INF/LICENSE content
org.apache.dao	None	Apache 2.0 only
org.apache.config	None	Apache 2.0 only
org.apache.auth	None	Apache 2.0 only
org.apache.pybuilder	None	Apache 2.0 only
org.apache.workflow-core	None	Apache 2.0 only
org.apache.workflow-operator	com/kjetland/jackson/jsonSchema/ (13 files, mbknor-jackson-jsonschema, MIT)	Apache 2.0 + MIT attribution
org.apache.config-service	None	Apache 2.0 only
org.apache.access-control-service	None	Apache 2.0 only
org.apache.computing-unit-managing-service	None	Apache 2.0 only
org.apache.file-service	None	Apache 2.0 only
org.apache.workflow-compiling-service	None	Apache 2.0 only
org.apache.amber (WorkflowExecutionService)	None	Apache 2.0 only

Priority

P1 – High

Task Type

• Code Implementation
• Documentation
• Refactor / Cleanup
• Testing / QA
• DevOps / Deployment