Hi Eric, when you say new instance, you mean instance created with instance:create or a "regular" vanilla one (maybe you don't use the right deploy folder) ?
You should have something in the log anyway from the artifact installer. Regards JB On 02/12/2018 21:15, Eric Lilja wrote: > Hi, see my replies at relevant places below! > > On 2018-12-02 21:03, Jean-Baptiste Onofré wrote: >> Hi Eric, >> >> The update to maven-bundle-plugin 4.1.0 introduces some other changes >> (as the requirements/capabilities created are extended). Why do you >> need this for 4.2.2 ? It's transparent for you as user. >> So I will try but without strong commitment. Reasonably, I would >> prefer to target in 4.2.3 as I'm planning 4.2.2 for this week. So, it >> will be a best effort. > > Okay, I see, I bumped to that version for my projects without issue so I > thought it was straight-forward. > >> >> Yes, deploy folder is enabled by default. Did you deploy a bundle or a >> raw jar ? > > It was one of my bundles, built using maven-bundle-plugin 4.1.0. I have > no troubles provisioning it programmatcially when using Karaf/Pax-Exam. > I also tried with a 3PP found on Maven central. The deploy folder > doesn't seem to be active on Windows. It was working fine on Windows a > couple of versions (of Karaf) ago. > >> >> The line breaks should be fixed already (you can test on 4.2.2-SNAPSHOT). > > Ah, nice to hear! Could you try on your end Windows with TakeCommand to > see about the question-mark-thingie? I also recall a DOS .bat-file error > being reported by TakeCommand as I launched karaf.bat > > - Eric L > >> >> Regards >> JB >> >> On 02/12/2018 20:15, Eric Lilja wrote: >>> Hi, is it possible to consider including KARAF-5934 for 4.2.2? I saw >>> it was just tentatively postponed to 4.2.3. >>> >>> Also, is the deploy folder enabled by default? I downloaded the >>> binary distribution (version 4.2.1), and created a new instance with >>> feature scr enabled. Everything else vanilla. I put a bundle in the >>> deploy folder but it doesn't get installed, no interesting output >>> from log:tail either, so no failures. Just like the directory was not >>> monitored at all (I did use the deploy directory in the instance I >>> created). >>> >>> Furthermore, output from scr:list seems to lack line breaks. This is >>> under Windows, with Cygwin (Mintty) as terminal. Also tried with >>> TakeCommand under Windows, which used to work really well with Karaf, >>> but when one does bundle:list -s -t 0 for example, a lot columns >>> contains only ?-characters (question marks). Also scr:list looks bad >>> under it. >>> >>> - Eric L >>> > -- Jean-Baptiste Onofré [email protected] http://blog.nanthrax.net Talend - http://www.talend.com
