Historically we have supported old baselines for a long time in the build toolchain. The current build toolchain supports baselines as far back as 2.249. This provides a high degree of flexibility for plugin maintainers and ultimately end users.
Sustaining this level of compatibility across complex transitions like the move to Java 11 and Jetty 10 comes at significant build toolchain complexity, either in the form of conditional logic in a single-branch build toolchain scheme (the current status quo) or in the form of multiple branches (a hypothetical scenario). For example, the test harness needs to support Java 8 in order to support 2.249, the need to support Java 8 means the test harness cannot move to Jetty 10 (which requires Java 11), and the test harness' continued use of Jetty 9 requires core to retain support both Jetty 9 and 10 (even though only Jetty 10 is used in production use cases). Similarly, conditional logic is required to set the Maven compiler release version to 11 for plugins when newer baselines are in use, and even then that conditional logic tends to trip up IDEs. For these reasons, it is desirable to require a baseline of 2.361 or later (and therefore Java 11) for plugins in the build toolchain sooner rather than later. The advantage is that this allows us to simplify and clean up conditional logic that is increasingly becoming less relevant and a maintenance burden. The disadvantage of requiring a baseline of 2.361 or later for plugins in late 2022 or early 2023 (as opposed to our usual timeline of e.g. late 2023 or early 2024) is that it is more of an imposition on plugin maintainers and users of older LTS lines. I believe that the advantages outweigh the disadvantages and that we should do this sooner rather than later, ideally in late 2022. If the consensus is that we are OK with requiring a baseline of 2.361 or later for plugins in late 2022 (i.e., after the release of 2.375.x LTS) or early 2023 (i.e., after the release of the next major LTS line after 2.375), then I explicitly commit to implement the change in the plugin parent POM (e.g., with regard to Java 8), test harness (e.g., with regard to Unicode properties files), HPI plugin (e.g., with regard to the Maven compiler release version), core (e.g., with regard to Jetty 9), annotation indexer, symbol annotation library, version number library, and access modifier library, noting in the release notes for all eight (8) components that this is a breaking change for plugin maintainers and that plugin maintainers must use a baseline of 2.361 or later when upgrading to the latest build toolchain. If you are in favor of requiring a baseline of 2.361 or later for plugins in late 2022 or early 2023 and with me implementing this transition in the manner described, please reply with your timeline preference (either late 2022 or early 2023). If you are in favor of maintaining a stable branch of the abovementioned eight (8) components for older LTS lines, please reply stating whether you are willing to do the implementation and release work (noting that I explicitly decline to do this implementation and release work). If you are in favor of any other proposal, please reply stating the alternative proposal and whether you are willing to do the implementation work. Thanks, Basil -- You received this message because you are subscribed to the Google Groups "Jenkins Developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to jenkinsci-dev+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/jenkinsci-dev/CAFwNDjo4AV%3DuY-f1R%3DjTFTfwe-m8TdxONt4Ly7yhU3JpPTJj6w%40mail.gmail.com.
