What should we do?
From my experience creating https://github.com/crobbins215/pdf-merge-split, could we investigate if this template can be improved
- Update README
- Add a decision guide: Operations API vs Classic Function API (when to use which).
- Add a 5‑minute Quickstart: “Use this template” → rename checklist → mvn clean package.
- Document local vs SaaS run paths (Local Connector Runtime vs Docker bundle), with Windows-friendly commands.
- Element template tips: typed defaults (no string booleans/numbers), prevent duplicate “Operation” fields, icon guidance.
- Troubleshooting: common pitfalls (wrong test imports, missing SPI provider, duplicate properties, version mismatches).
- Web Modeler testing vs Desktop, these examples can be improved.
- Add a document-handling example (new examples/document-handling/)
- Show both in-memory bytes and stream-based handling with the Connector Document API.
- Include large file guidance (heap notes, temp files, streaming, max size), metadata handling, and multiple documents pattern.
- Provide unit tests that demonstrate processing small and large documents safely.
- Provide a minimal, working test using connector-runtime-test
- Example test with OutboundConnectorContextBuilder (SDK 8.8.3).
- One happy-path test that compiles and passes out-of-the-box.
- Add an element-template “do/don’t” section
- Unique property IDs, grouping, hiding discriminators, typed defaults, examples.
- Optional: add a tiny validator script/goal or checklist to catch duplicate IDs and bad defaults before opening Modeler.
- Pin and surface versions
- SDK and generator plugin pinned to latest stable (currently 8.8.3).
- Note a simple compatibility matrix (SDK, Modeler, runtime, test libs).
Why should we do it?
- Reduce time-to-first-success for partners and customers (target: <10 minutes from clone to passing test and a clean element template in Modeler).
- Eliminate recurring pitfalls:
- Duplicate “Operation” fields from discriminator leaks.
- Mis-typed defaults in templates (booleans/numbers as strings).
- Wrong test imports (connector-test vs connector-runtime-test).
- Confusion running locally vs SaaS. Webmodeler vs desktop modeler
- Lower support load and improve community connector quality by providing a clear, copy‑pasteable baseline.
- Encourage correct use of the Document API with practical guidance for large files and memory.
- Align template with current SDK versions and best practices, improving consistency across community contributions.
What should we do?
From my experience creating https://github.com/crobbins215/pdf-merge-split, could we investigate if this template can be improved
Why should we do it?