Fred Bricon
Fred Bricon
Upstream bug: https://github.com/eclipse-jdt/eclipse.jdt.core/issues/4623
@snjeza before you try to fix m2e, do you have a general idea on how to do it?
Mmm that would not work OOTB. And you probably need to handle javadoc and sources differently. An alternative might be to, before attaching the file, inspect the 1st (java) file...
@snjeza If going with ICU4J, make sure you're not depending on the icu4j bundle, but the packages, since jdt.ls only embeds the [base icu4j jar](https://github.com/eclipse/eclipse.jdt.ls/issues/252)
@snjeza I'm curious, does icu4j detect anything other than utf-8 if you scan the entire junit-3.8.1-sources.jar?
- how long does it take to full-scan a jar? - I'm interested to know whether we can accurately detect encoding after scanning say 5, 10, 100 files in a...
Since encoding detection gives worse results than the default UTF-8 encoding, we're gonna need a new command to explicitly set encoding to a URI (now that m2e persists that information).
WIP to expose a setSourceEncoding command: 
Yeah I basically told the same thing to @jjohnstn ;-)
We'll stick with 4.38 GA until the release next week.