Ok, part way through writing this email I realized it probably makes some sense to break it down a bit more. Now I'm thinking we need an AnnotationProcessor interface with methods like: String getClassName() String getOptions() String getSourceDirectory() String getOutputDirectory() List<Dependency> getDependencies() Then the AnnotationProcessorFacet methods would be: addAnnotationProcessor(AnnotationProcessor processor) removeAnnotationProcessor(AnnotationProcessor processor) List<AnnotationProcessor> getAnnotationProcessors() processAnnotations() processAnnotation(AnnotationProcessor processor) That's some real thinking on the fly so any other thoughts/enhancements are most welcome! Ken On Wed, Jul 6, 2011 at 5:04 PM, Lincoln Baxter, III < lincolnbaxter(a)gmail.com&gt; wrote: > Hmm, > > Good questions. I think #2 would probably be the easiest, but we could > offer #1 as an alternative for those who don't want to use the maven build > plugin. > > I think #3 would probably be the most robust for future use. What kinds of > methods are you thinking of including for the facet? > > ~Lincoln > > > On Sun, Jul 3, 2011 at 3:41 PM, Ken Finnigan <ken(a)kenfinnigan.me&gt; wrote: > >> Hey Lincoln, >> >> Have been working on adding JPA model generation to the persistence >> plugin over the last few days and have been mulling over the best way to >> implement it and wanted to get your thoughts. >> >> I see that there are a few possibilities: >> >> 1) Call the NativeMethodExec utility to call javac for running annotation >> processing only. >> 2) Add an annotation processing profile to the pom as part of the install >> of the plugin (if they select they want model gen), and then performing the >> model gen would execute the maven to process and generate annotations >> 3) An extension of 2 would be to make an AnnotationProcessorFacet, and >> then do 2 as a Maven implementation of that facet. >> >> After thinking about it today I think 3 seems the most flexible, but >> wanted to get your thoughts. >> >> Thanks >> Ken >> > > > > -- > Lincoln Baxter, III > http://ocpsoft.com > http://scrumshark.com > "Keep it Simple" >

In a Maven sense I'm pretty sure they run in compile phase, though there is also a Maven plugin I've found to run annotation processors directly, so we could take either approach. On Thu, Jul 7, 2011 at 10:08 AM, Lincoln Baxter, III < lincolnbaxter(a)gmail.com&gt; wrote: > Forgive me for asking something I could easily google, but where do > annotation processors fit into the development lifecycle? I admit I've never > used one (that I know of.) It would probably help if I did before giving too > much feedback ;) > > ~Lincoln > > > On Thu, Jul 7, 2011 at 9:12 AM, Ken Finnigan <ken(a)kenfinnigan.me&gt; wrote: > >> Ok, part way through writing this email I realized it probably makes some >> sense to break it down a bit more. >> >> Now I'm thinking we need an AnnotationProcessor interface with methods >> like: >> >> String getClassName() >> String getOptions() >> String getSourceDirectory() >> String getOutputDirectory() >> List<Dependency> getDependencies() >> >> >> Then the AnnotationProcessorFacet methods would be: >> >> addAnnotationProcessor(AnnotationProcessor processor) >> removeAnnotationProcessor(AnnotationProcessor processor) >> List<AnnotationProcessor> getAnnotationProcessors() >> processAnnotations() >> processAnnotation(AnnotationProcessor processor) >> >> That's some real thinking on the fly so any other thoughts/enhancements >> are most welcome! >> >> Ken >> >> >> >> On Wed, Jul 6, 2011 at 5:04 PM, Lincoln Baxter, III < >> lincolnbaxter(a)gmail.com&gt; wrote: >> >>> Hmm, >>> >>> Good questions. I think #2 would probably be the easiest, but we could >>> offer #1 as an alternative for those who don't want to use the maven build >>> plugin. >>> >>> I think #3 would probably be the most robust for future use. What kinds >>> of methods are you thinking of including for the facet? >>> >>> ~Lincoln >>> >>> >>> On Sun, Jul 3, 2011 at 3:41 PM, Ken Finnigan <ken(a)kenfinnigan.me&gt;wrote: >>> >>>> Hey Lincoln, >>>> >>>> Have been working on adding JPA model generation to the persistence >>>> plugin over the last few days and have been mulling over the best way to >>>> implement it and wanted to get your thoughts. >>>> >>>> I see that there are a few possibilities: >>>> >>>> 1) Call the NativeMethodExec utility to call javac for running >>>> annotation processing only. >>>> 2) Add an annotation processing profile to the pom as part of the >>>> install of the plugin (if they select they want model gen), and then >>>> performing the model gen would execute the maven to process and generate >>>> annotations >>>> 3) An extension of 2 would be to make an AnnotationProcessorFacet, and >>>> then do 2 as a Maven implementation of that facet. >>>> >>>> After thinking about it today I think 3 seems the most flexible, but >>>> wanted to get your thoughts. >>>> >>>> Thanks >>>> Ken >>>> >>> >>> >>> >>> -- >>> Lincoln Baxter, III >>> http://ocpsoft.com >>> http://scrumshark.com >>> "Keep it Simple" >>> >> >> > > > -- > Lincoln Baxter, III > http://ocpsoft.com > http://scrumshark.com > "Keep it Simple" >