> In a way I am honored that your "hard earned experience" with OSGi and > Maven's BND plugin, is all blamed on the OPS4J community behind the > Pax toolkit. We are hardly responsible for "everything OSGi"...
It's AFU that my writing style comes across as blaming. It's a genuine point of inquiry for me, and indicates a personal defect that instead it comes across as creepy blaming behavior. To the OPS4J community I represent the mainstream corporate developer who only wants to consume OSGi but not understand it. In my case, pax represents the ability to start with something that works, and then adjust it until I begin to understand what is going on under the covers. That's a really stupid way to learn OSGi, but I didn't know that until I tried it and failed. It may be off topic, but I still maintain that pax can bring OSGi into the mainstream, and that our industry wins when this happens, even if it is too big a subject for this thread. After posting this note I realized my expectations were out of line. I consider it my job to clarify things in the near future, so that the idiot corporate developer like me who comes in not wanting to understand OSGi but just to do it, can use pax with some degree of success and not waste the time that I did. That's only fair. I should have something up within a few weeks that documents a reasonable approach from what exists in pax. Hopefully by then my writing style will not come across so creepy like. > > But I hope you feel better. > > And if you do; Could you please outline the workflow you performed > (esp docs referred to) to get from 'nothing' to 'problems'. I suspect > that some Eclipse PDE docs have been interfering, because if you at > all got to BND docs, then you should be pretty clear that it is a > macro processor for generation of the manifest file. > > On Thu, Dec 9, 2010 at 1:05 AM, Pete Carapetyan > <[email protected]> wrote: >> Wondering if my experience below is really what the pax community >> intends for the [hundreds, thousands of potential?] newbie OSGi >> developers who might wish to use pax as it's entry point to the world >> of OSGi? >> >> Per earlier, I'm an OSGi newbie attempting to learn and update docs >> for pax OSGi to work in a Maven/Spring/OSGi combination. I have many >> hours trying to learn why my pax generated projects don't work when I >> run them in pax exam. Very time consuming. >> >> To make a very long story short - I write a MANIFEST.MF and Apache >> Maven Bundle Plugin re-writes it during mvn install to suit itself. >> Which would be fine if >> >> a. I had known it was happening and knew to check it before burning >> hour after hour checking everything else that didn't matter. >> b. The plugin itself overwrote my MANIFEST.MF correctly >> >> Bottom line is that it strips out some of the packages I need from my >> Import-Package: statements that I had very explicitly written. Once I >> learned that it was doing this, 15 minutes later and I've got >> everything fixed using the osgi.bnd file. Nevermind the hours and >> hours spent looking everywhere but here to figure out what was going >> wrong. Basically if you don't unzip and examine the generated >> Import-Package: you never even know that you got pimped. >> >> Something seems horribly wrong here if we intend to attract the new >> developers that OSGi would require to hit critical mass and thus >> support pax's work. Am I all washed up on this? Comments? >> >> _______________________________________________ >> general mailing list >> [email protected] >> http://lists.ops4j.org/mailman/listinfo/general >> > > > > -- > Niclas Hedhman, Software Developer > http://www.qi4j.org - New Energy for Java > > I live here; http://tinyurl.com/3xugrbk > I work here; http://tinyurl.com/24svnvk > I relax here; http://tinyurl.com/2cgsug > _______________________________________________ general mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/general
