GitHub user sramazzina added a comment to the discussion: 2.17 Roadmap and goals
Hi Hans and community Here are our two cents on this post. I'm using the plural because the following suggestions/desiderata comes from myself and my team who work every day with Hop, both for our customers and as contributors. Let me say that, frankly, we consider the UI the very last of the problems in Hop at the moment. >From our experiences with our clients (and the issues we have encountered), >the main point in the TODO list is not the UI, but improving code reliability >and trust. Therefore, we strongly suggest starting by working in these areas: ### Stability - **UI tests** – the introduction of SWTBot for testing automation (or other alternatives) would definitely bring a massive impact: lot of issues reported by our clients in these years were about the UI; this would help out to prevent regressions; also, it will make us ways more confident in case of refactoring. - **Add more integration tests**, even if they require a long execution time. If they take too long, we must find a good balance. We need to avoid a decline in product quality, as seen in the last two versions, and manage any issues as soon as they arise. - **Run integration tests after merging.** Maybe not every time we merge a PR, because they take a long time to execute, but perhaps once a day or periodically. ### PR policies - **PRs must be atomic** One issue per single PR must be the mandatory approach, and not many issues in a single PR, as sometimes happens now. This ensures easier code reviews, clearer traceability of changes, and faster identification of the cause in case of regressions. It also simplifies reverting problematic commits and improves overall project stability. - **Merge** only if the PR has a test that confirms the change is useful - **Code review.** We think it would be best practice to start requiring more than one reviewer for a PR. No review, no merge d PR. Then we can think at the rest that, shortly, we can summarize below ### Features - **Finalize ongoing developments.** In addition to the points above, specifically related to increasing the reliability of our distribution, we believe it is necessary to complete some developments that have remained unfinished for years. I am referring, for example, to the completion of the XML cleanup in favor of annotations (HopMetadataProperty), so that we can save pipeline/workflow metadata using other file types, such as JSON or YAML. We have more than three pages of open issues, such as “XML cleanup”, that no one is looking at right now (including me and my team). - **Project Inheritance.** @hanva, you disagree on this, and you said you don't see any real application in production use cases. We can talk about that. I only said that it is something that needs to be definitely completed because, as it is now, it is completely useless. Better to clean up the code and remove it completely IMHO. As we notice so far, part of the desiderata concerns code, and part concerns a change in the way we manage the project. Of course, we are well aware that these issues will not be resolved in the next version, but we wanted to give a concrete vision — albeit a long-term one (let's say a year from now) — for our beloved Hop. Cheers GitHub link: https://github.com/apache/hop/discussions/5980#discussioncomment-14947267 ---- This is an automatically sent email for [email protected]. To unsubscribe, please send an email to: [email protected]
