apache / incubator-kie-issues

Apache License 2.0
11 stars 1 forks source link

Stunner: Differences assessment of the Stunner codebases (BPMN/DMN/SCESIM vs SWF) #47

Closed yesamer closed 1 year ago

yesamer commented 1 year ago

As a result of the project management changes announced 10 months ago, the Stunner codebase has been forked in two different streams:

  1. BPMN/DMN/SCESIM stream
  2. SWF (Serverless Workflow) stream

In the last months, several bug fixing, improvements, and new features have been merged in stream 2. We need to evaluate which of those changes are worthy to be "cherrypicked" in stream 1.

jomarko commented 1 year ago

@yesamer to me it seems, we have in codebase now 3 clones of appformer/uberfire. Is addressing it somehow part of #47 too?

Screenshot from 2022-11-30 11-38-29

yesamer commented 1 year ago

@jomarko Yep good point. I would suggest ignoring the Dashbuilder fork because I consider it a totally separate project from ours. In addition, I don't see any valuable fix/feature in its stunner modules (https://github.com/kiegroup/kie-tools/commits/main/packages/dashbuilder/appformer), these are mostly dependencies for security updates and Dashbuilder-specific issues. WDYT?

jomarko commented 1 year ago

My point was rather, aren't we building the same three times during CI for example?

yesamer commented 1 year ago

@jomarko Our CI/CD is smart enough to build the modules related to PR changes. Eg. if you modify something on DMN Editor module, the CI will build its parent modules only, and not all the modules in kie-tools. Please correct me if I'm wrong @tiagobento

yesamer commented 1 year ago

I completed my analysis (with @romartin support, thank you, man!)

Fix and Performances fixes to "cherrypick" (Higher Priority)

Nice to have new features (Lower Priority)

All other PRs are related to SWF only.

The tickets of the first list can be cherrypicked as a 1:1 copy. The tickets of the second list require integration with the editors, so the required effort is higher.

For this ticket scope, my proposal is to cherrypick the items of the first list only

yesamer commented 1 year ago

KOGITO-6974 - DONE KOGITO-7736 - Discarded: Performance improvements are not significant in BPMN/DMN/SCESIM stream. Please check https://github.com/kiegroup/kie-tools/pull/1357 for further consideration.