Guillaume- I know this to be true (and have seen it with my own eyes), but I can't find where it's documented. Can you send me a link?
In any case, I agree with you that a mandatory package shouldn't be created. Ideally, an optional one would be created for RUNTIME or CLASS, but none for SOURCE. Thanks, Justin On 11/19/10 11:05 AM, Guillaume Nodet wrote: > Note that loading a class with annotations while annotations are not > available never result in a NoClassDefFoundError, annotations are > simply discarded, so I'm not sure that a mandatory package should be > generated, maybe an optional one though. > > On Fri, Nov 19, 2010 at 16:21, Justin Edelson <jus...@justinedelson.com> > wrote: >> Peter- >> This is perhaps a bit of lazymail, but will bnd by default create an >> Import-Package statement for CLASS-retained annotations? >> >> In other words, if you had >> >> package foo; >> >> import bar.MyAnnotation; >> >> @MyAnnotation >> public class MyClass { >> >> } >> >> Where bar.MyAnnotation has a RetentionPolicy of CLASS and is in a separate >> bundle, what Import-Package header for the bundle containing package foo >> will be generated by default? >> >> Thanks, >> Justin >> >> >> On Nov 19, 2010, at 8:58 AM, Peter Kriens wrote: >> >>> There is a class in bnd that can do this: aQute.lib.osgi.Clazz, it does not >>> need access to the actual annotation classes. >>> >>> bnd uses CLASS annotations for its DS support because it is a lot easier to >>> parse byte codes than source code. >>> >>> Kind regards, >>> >>> Peter Kriens >>> >>> >>> On 18 nov 2010, at 20:47, Felix Meschberger wrote: >>> >>>> Hi, >>>> >>>> Am Donnerstag, den 18.11.2010, 11:50 -0500 schrieb Justin Edelson: >>>>> On Thu, Nov 18, 2010 at 3:35 AM, Felix Meschberger <fmesc...@gmail.com> >>>>> wrote: >>>>>> Hi, >>>>>> >>>>>> Am Sonntag, den 01.08.2010, 14:07 +0200 schrieb Reto Bachmann-Gmuer: >>>>>>> Hi Atle >>>>>>> >>>>>>> In the clerezza projects we're using more and more scala, the biggest >>>>>>> disandvantage for me is that the maven-scr-plugin doesn't work for >>>>>>> scala so >>>>>>> I've to write the ds component descriptor by hand. We generallay don't >>>>>>> use >>>>>>> objects but classes, having ds caring about creating and activating the >>>>>>> instances. >>>>>> >>>>>> I am by no means a Scala expert (whatever you will say now, but I am >>>>>> just scared by the syntax ;-) ) >>>>>> >>>>>> So, coming back to the scr-plugin problem: The plugin (currently) reads >>>>>> the source files with the help of the QDox library. So I would assume >>>>>> that the inability of the plugin to process Scala is related to this >>>>>> situation. >>>>>> >>>>>> If you would know of a tool to read Scala source files for consumption >>>>>> by the plugin, you are welcome to guide us there (or even better provide >>>>>> a patch to use it ;-) ). >>>>>> >>>>>> I think a similar problem exists for Groovy (and yes, it would be nice >>>>>> to have something there, too ;-) ). >>>>> >>>>> In theory, you might be able to use annotations (i.e. "real" >>>>> annotations, not JavaDoc/QDox annotations). I've been planning on >>>>> trying this technique with Groovy and maven-scr-plugin, but I haven't >>>>> had the time to try it out. Beyond the fact that I personally prefer >>>>> to use @Component instead of @scr.component, it just like parsing >>>>> Scala (or Groovy) sources is far more complex than using the >>>>> reflection API. To be clear, this is still going to require surgery to >>>>> the scrplugin code, I just have a feeling that the surgery is going to >>>>> more like foot surgery than brain surgery (if this metaphor makes >>>>> sense). >>>> >>>> The point about having QDox also parse for the Annotations is, that the >>>> Annotations are defined to not be added to the class files to not create >>>> run-time dependencies and to not create class-file incompatibilities. >>>> >>>> Now, you may say we were over-cautious in this respect and another >>>> retention policy would be viable (in terms of not creating runtime >>>> dependencies), e.g. CLASS or even RUNTIME. >>>> >>>> If we can go with that, it should -- theoretically -- be possible to >>>> actually read the annotations from the classes instead of using the QDox >>>> parser. >>>> >>>> Regards >>>> Felix >>>> >>>>> >>>>> Justin >>>>> >>>>>> >>>>>> Regards >>>>>> Felix >>>>>> >>>>>>> >>>>>>> >>>>>>> Cheers, >>>>>>> reto >>>>>>> >>>>>>> On Sun, Aug 1, 2010 at 8:37 AM, Atle Prange <atle.pra...@gmail.com> >>>>>>> wrote: >>>>>>> >>>>>>>> Thank you for the replies, my journey can continue, although with a >>>>>>>> sligth >>>>>>>> blow to my self esteem, i thought i was an experienced programmer, i >>>>>>>> guess >>>>>>>> i >>>>>>>> have to read more books.... >>>>>>>> >>>>>>>> >>>>>>>> -atle >>>>>>>> >>>>>>>> On Sat, Jul 31, 2010 at 11:02 PM, Christopher Brind >>>>>>>> <bri...@brindy.org.uk >>>>>>>>> wrote: >>>>>>>> >>>>>>>>> Have never used Scala myself, but Peter Kriens discussed this in his >>>>>>>>> blog >>>>>>>>> recently: >>>>>>>>> http://www.osgi.org/blog/2010/07/scala-components-vs-osgi.html >>>>>>>>> >>>>>>>>> Hope it helps in some way. >>>>>>>>> >>>>>>>>> Cheers, >>>>>>>>> Chris >>>>>>>>> >>>>>>>>> >>>>>>>>> On 31 July 2010 21:58, Atle Prange <atle.pra...@gmail.com> wrote: >>>>>>>>> >>>>>>>>>> Hi, >>>>>>>>>> >>>>>>>>>> i am about to embark on a journey involving scala and osgi. But it >>>>>>>>> appears >>>>>>>>>> to me (ref. my last question on this list regarding static >>>>>>>>>> references) >>>>>>>>> that >>>>>>>>>> it might not be a good idea, since the otion of scala objects >>>>>>>>>> actually >>>>>>>>> are >>>>>>>>>> implemented as singletons, and therefor never will be cleaned up >>>>>>>>>> after >>>>>>>> a >>>>>>>>>> bundle is unloaded. Does anybody here have experience with scala on >>>>>>>> osgi, >>>>>>>>>> and could give me a hint on this matter? >>>>>>>>>> >>>>>>>>>> -atle >>>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> --------------------------------------------------------------------- >>>>>> To unsubscribe, e-mail: users-unsubscr...@felix.apache.org >>>>>> For additional commands, e-mail: users-h...@felix.apache.org >>>>>> >>>>>> >>>>> >>>>> --------------------------------------------------------------------- >>>>> To unsubscribe, e-mail: users-unsubscr...@felix.apache.org >>>>> For additional commands, e-mail: users-h...@felix.apache.org >>>>> >>>> >>>> >>>> >>>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: users-unsubscr...@felix.apache.org >>>> For additional commands, e-mail: users-h...@felix.apache.org >>>> >>> >>> >>> --------------------------------------------------------------------- >>> To unsubscribe, e-mail: users-unsubscr...@felix.apache.org >>> For additional commands, e-mail: users-h...@felix.apache.org >>> >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: users-unsubscr...@felix.apache.org >> For additional commands, e-mail: users-h...@felix.apache.org >> >> > > > --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@felix.apache.org For additional commands, e-mail: users-h...@felix.apache.org