Hi Alex,
as you seem to be responding to everything else I guess you missed my asking you to rollback your changes and move them to a feature branch. Are you going to revert b8d8c48b11f41ea435eefafc94eb0927229e8a4b It's ok that it "slipped out accidentally", but it's not ok to keep things broken for so long. Having a look at the change-set, this was quite a large commit and I will have to do a little digging into the problem to find out what you changed. Also writing a documentation that will help someone not really working with Maven to understand how to fix the problem will take some time. But I've heard writing core documentation is a task left to people feeling an itch and that time is better invested in producing features (Sorry ... couldn't resist ... but maybe it's worth mentioning as exactly this sort of single-point-of-failure seems to exist on a lot of the compiler from my point of view. I bet most people on the list haven't got a clue at what's happening in the compiler ad we don't have documentation for that either) Unfortunately I currently don't have that much time to do the fixing myself or write that documentation immediately and I would prefer a little discussion on what's changed and not having to reverse engineer mentioned commit. Chris ________________________________ Von: Christofer Dutz <christofer.d...@c-ware.de> Gesendet: Dienstag, 4. Oktober 2016 08:35:33 An: dev@flex.apache.org Betreff: AW: Jenkins still red I agree ... the Plugin is not broken, but the compiler changed without the plugin having the chance to be adjusted to those changes. Yes most of the build logic is in the plugin itself. That's the way things are done in Maven. If you change the way things are built, you have to adjust the plugin to reflect those changes. I think it would be best, if you revert your changes and re-apply them in a feature-branch. If you name it "fleature-autobuild/*" it will automatically be built by the ASF Jenkins and we will be able to see when it's finished and can be merged back. In the meanwhile, I'll add some documentation about the plugin to the compilers site. Chris ________________________________ Von: Alex Harui <aha...@adobe.com> Gesendet: Montag, 3. Oktober 2016 20:47:58 An: dev@flex.apache.org Betreff: Re: Jenkins still red On 10/3/16, 11:41 AM, "carlos.rov...@gmail.com on behalf of Carlos Rovira" <carlos.rov...@gmail.com on behalf of carlos.rov...@codeoscopic.com> wrote: >Hi Alex, > >I think what I propose is better to do the next time. Right now better to >finish the ticket. > >Regarding the maven plugin, as I understand, is not broken...there's >nothing to fix, as I understand is a problem os code broken. If not I'm >missing something and better Chris to throw some light on this It appears the plugin might make an assumption that it has to run COMPJSC before COMPC and we need to reverse the order and probably change the compile-xx-config files and maybe the -output parameter sent to COMPJSC. I'm a Maven newbie. I was hoping to find both tools specified in the <build> section of the pom.xml but I didn't see it there, so now I'm thinking the plugin sets the order, but I couldn't find how that worked when looking through the plugin source. I'm sure I'm missing something somewhere. Next time, I hope to know enough about Maven and the plugin that I can get it all working locally before pushing. -Alex
