Hi Dan, Thanks for the inputs.
The annotation is already RetentionPolicy.RUNTIME. Current the html generator is based the APT processor, But it looks like JDK6 has some trouble to load the APT processor when compile the code, I’m not sure if I can fix it tomorrow. At mean while, I suggest we can spend some time to address this JDK compile issue after Camel 2.12.3 is released, so it won’t block the ActiveMQ release. -- Willem Jiang Red Hat, Inc. Web: http://www.redhat.com Blog: http://willemjiang.blogspot.com(http://willemjiang.blogspot.com/) (English) http://jnn.iteye.com(http://jnn.javaeye.com/) (Chinese) Twitter: willemjiang Weibo: 姜宁willem On February 20, 2014 at 10:29:47 PM, Daniel Kulp (dk...@apache.org) wrote: > > > On Feb 20, 2014, at 1:45 AM, Jan Matèrne (jhm) > wrote: > >> I'll give it another 12 hours and if nothing changes I will close > >> the vote and release 2.12.3. If within this time you convince > another > >> PMC member to change his vote I will cancel this vote. > > > > I am also a fan a building against the oldest supported platform. > > Would it be possible to > > - build Camel against Java6 for getting the "right" class file > format > > - generate the html using Java7 > > - jar them using Java6? > > We could easily make the annotations be RetentionPolicy.RUNTIME > so they are kept in the compiled classes. At that point, we have > two options that could work with java6: > > 1) Write a utility class that would take a Class object and generate > the appropriate HTML. Tools could easily use this. This could > be designed with protected methods or something so that they > could override parts of it to customize what is generated as they > need. > > 2) Create a maven plugin that runs in process-classes to generate > static HTML. It could possibly even use the utility from (1). > > Again, both would work fine with java 6. > > In any case, I also strongly believe that anything that is NOT > built as part of using our lowest supported JDK is “optional”. > If it is or is not part of the convenience artifacts is irrelevant > from a release standpoint. If people want that functionality, > find a way to make it not optional. > > > Dan > > > > > Another option would drop Java6 ... But as Oracle writes [2] > for Java6: > > - Feb 2011: End of Public Updates Notification > > - Feb 2013: End of Public Updates > > - Dec 2013: Premier Support Until > > BUT > > - Jun 2017: Extended Support Until > > > > > > [2] http://www.oracle.com/technetwork/java/eol-135779.html > > > > Jan > > > > -- > Daniel Kulp > dk...@apache.org - http://dankulp.com/blog > Talend Community Coder - http://coders.talend.com > >
