09:13 <nowherefast[m]> Spent some time this weekend on a way overdue refresh of a couple of jruby-javafx apps, jdk 8 -> jdk 17. Besides JRubyFX off the main branch and switching ant out for jpackage, everything else just worked. 🙇

09:19 <nowherefast[m]> One quirk that might be mine to solve has to do with `require_relative` in files that have been put through **jrubyc**. This worked fine previously (and still does when the require is in rb files), but lands me a complaint when in compiled files:... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/e20826699bfef82ef1ab5eaef0fd851e0bf69f13)

09:24 <nowherefast[m]> Seems to be looking for the file on the system than in the classloader namespace. Do I need to be explicit in some way with the `require_relative` in files that will be compiled?

09:33 <headius> If I'm remembering right we removed the search for class files because it basically doubled the overhead of every require

09:33 <headius> There is a flag you can specify that will search for them again

09:34 <headius> Oh but you were saying the compiled files don't do the relative require correctly

09:34 <headius> Yeah that could be a bug

09:34 <headius> nowherefast:

09:35 <nowherefast[m]> yup, if I don't put them through compile, it all works, all the way through to running off the jar and after packaging.

10:36 <headius> Yeah that's worth an issue

20:08 <byteit101[m]> nowherefast: "JRubyFX off the main branch" Oh good, that works! there are some bugs that I need to get around to fixing in FXML components, but good to hear someone is getting it to work