Fair enough, done in https://bz.apache.org/bugzilla/show_bug.cgi?id=62986
Romain Manni-Bucau @rmannibucau <https://twitter.com/rmannibucau> | Blog <https://rmannibucau.metawerx.net/> | Old Blog <http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau> | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book <https://www.packtpub.com/application-development/java-ee-8-high-performance> Le jeu. 6 déc. 2018 à 14:32, Mark Thomas <ma...@apache.org> a écrit : > On 06/12/2018 13:19, Romain Manni-Bucau wrote: > > Hello Rémy, > > > > > > Le jeu. 6 déc. 2018 à 14:13, Rémy Maucherat <r...@apache.org> a écrit : > > > >> On Thu, Dec 6, 2018 at 11:51 AM Romain Manni-Bucau < > rmannibu...@gmail.com> > >> wrote: > >> > >>> Hi guys, > >>> > >>> can you make ContextConfig.JavaClassCacheEntry public please? Idea is > to > >> be > >>> able to override ContextConfig and potentially customize > >> processAnnotations > >>> methods. Currently it is a pain and it is preventing to upgrade > >> meecrowave > >>> and likely tomee to java11+jlink support > >>> > >> > >> Ok, so I had a look at the meecrowave main class after doing my embedded > >> refactoring and I was a bit horrified by the amount of hacks :) Not that > >> there were too many options to do the things that were done (some I > didn't > >> think were possible), and luckily my changes would have been a very good > >> help there. > >> > >> So here, it looks like a bit scary as well, since JavaClassCacheEntry is > >> package private (not ok usually), but only instantiated from a private > >> method (what would you do about that ?). So I'd rather not do that, or > not > >> just that, since it would be good to reduce the amount of hacks. > >> > >> For starters, I don't like "private" in that kind of class (same with > >> "final"), I prefer "protected" usually. Shouldn't JavaClassCacheEntry > be a > >> non static protected class ? > > No objections to making it protected if that helps but I don't see why > you'd want to make it non-static. > > >> About webConfig, I'm not sure. A good way to do it would be to add > calls to > >> a few intermediate empty methods in appropriate locations, rather than a > >> real refactoring. > > > > not empty cause here the idea is to reuse an existing scanning and not > let > > tomcat scan classes but yes a doScanMagic() would work > > the part which would be great to reuse is the mapping between bytecode > (the > > annotation entry) and the tomcat model (FilterDef etc) > > Why don't you suggest a patch? As a rough guideline, please aim for the > minimally invasive patch the gives you what you need. > > Generally adding to the existing API is OK but changing it is not. The > aim is to minimise the chance of breakage for anyone currently using or > extending that class directly. > > Mark > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > For additional commands, e-mail: dev-h...@tomcat.apache.org > >