Hey people, I was wrapping my head around ezbake and need some thoughts. ezbake is used as a build framework to glue the clojure/ruby/jruby/configs together.
• Right now when we build openvoxserver/db, we pull in the main branch from ezbake.:
https://github.com/OpenVoxProject/openvox-server/blob/main/tasks/build.rake#L67-L69
• that makes builds hard to reproduce and debug. If the actual openvoxdb/server code didn't change over time, but we rerun the last build, and ezbake changed in between, we get a different package
• The best way to test ezbake changes is to kick off builds for openvoxserver/db. So having an easy way to pass a git ref or rc release is helpful
• in openvoxserver/db, an ezbake version is also hardcoded:
https://github.com/OpenVoxProject/openvox-server/blob/main/project.clj#L174
• Ideally we version ezbake (
https://github.com/OpenVoxProject/ezbake/blob/main/project.clj#L1), but then openvoxserver/db also need to be updated to pull in those new versions
• perforce published their version to clojars:
https://clojars.org/puppetlabs/lein-ezbake
• We could also publish to clojars, including -snapshot versions. And then update those snapshot versions for each PR maybe?
• IMO: Ideally we can kick of openvoxserver/db builds, automatically, for each ezbake PR if we want to, before we merge it. But my knowledge about leiningen/clojars is limited.