Factor execution time and construction time dependencies into a non-SDK component
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 2d 2h
- Merged PRs (30d)
- 205
Description
Generally an execution engine should not depend on SDK constructs (e.g. DoFn), as those should be used by the language's SDK harness to execute a UDF. However, there are certain constructs that are used to communicate between the Runner and SDK harness, or are required to fulfill the Runner's responsibilities (e.g. Coder, Window). Those should be accessible without having to depend on the entire SDK
Imported from Jira [BEAM-3100](https://issues.apache.org/jira/browse/BEAM-3100). Original Jira may contain additional context.
Reported by: tgroh.
Contributor guide
Research direction
Start by reading the issue and its linked BEAM-3100 Jira entry to recover the missing context and identify the execution-time and construction-time dependencies. Map which Runner/SDK-harness communication constructs, especially Coder and Window, are required without the full SDK. Done means those constructs are accessible from a non-SDK component while execution engines no longer depend on SDK constructs such as DoFn.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- java
- Domain
- backend, distributed-systems
- Issue type
- Refactor
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100