Hi Chris, I convinced myself this just made sense so I committed the changes :-) I can back them out if you think groupId-per-platform is a mistake.
With these changes also comes the opportunity to make the j8,j7,android poms more closely resemble each other and also changes that may lessen the likelihood of copy/paste errors and such down the road. I’ve got another batch of changes in my workspace for these that I’ll likely deliver soon too. I have a dream that some day there may be some tooling that helps with managing the duplicate stuff resulting from the duplicate pom structure for each platform :-) — Dale > On Oct 18, 2017, at 9:22 AM, Dale LaBossiere <dml.apa...@gmail.com> wrote: > > Chris, > > I’ve got everything working in my workspace and I’ll commit it once you > verify the switch is desirable. > > (got the distribution stuff working — the include/exclude decls support GAV > pattern matching) > > — Dale > >>> On Oct 17, 2017, at 8:54 PM, Dale LaBossiere <dml.apa...@gmail.com >>> <mailto:dml.apa...@gmail.com>> wrote: >>> >>> FWIW, In my workspace I’ve made the changes to switch to a single >>> groupId-per-platform and so far I haven’t run into any issues. >>> I’ll report a full conclusion after exercising everything. >>> — Dale >>> >>>> On Oct 17, 2017, at 8:37 PM, Dale LaBossiere <dml.apa...@gmail.com >>>> <mailto:dml.apa...@gmail.com>> wrote: >>>> >>>> Hi Chris, >>>> >>>> Was it just arbitrary or is there an important reason why multiple maven >>>> artifact coordinate groupIds were defined? >>>> e.g., o.a.edgent.providers, o.a.e.connectors, o.a.e.api, etc >>>> Why not just o.a.e, o.a.e.java7, o.a.e.android? >>>> >>>> Browsing nexus (https://repository.apache.org >>>> <https://repository.apache.org/>) for a quick selection of projects >>>> (kafka, nifi, beam), I see they only have a single groupId. >>>> >>>> — Dale >>> >> >