Here's a sample script that reproduces this, since it wasn't immediately obvious from the report:

require 'kconv'

"test".isutf8

This appears to be the same issue as #338.

@headius the same issue as a three year old bug? 😿

Yeah...we need to fix NKF to define those constants as Encoding instances and do the mapping back and forth to NKF encoding ints.

@headius the above sample script doesn't reproduce. Works for me.

my env

$ ruby -v
jruby 9.0.5.0-SNAPSHOT (2.2.3) 2016-01-22 1761e2e Java HotSpot(TM) 64-Bit Server VM 25.66-b17 on 1.8.0_66-b17 +jit [darwin-x86_64]

@k77ch7 Hmm, you're right. I'll poke around a bit and see if there's any reason not to close this.

Ah-ha, our good friend @kares is responsible for fixing this in cfdccd7.

I'm removing the 1.8 paths for JRuby 9.1.0.0 but otherwise we'll call this fixed. I'll also investigate whether we've got some tests to untag/unexclude.

I removed the 1.8 paths in ecae248.

There does not appear to be any tests for the kconv.rb functionality right now.

Ok scratch that...there's nkf/test_kconv.rb. I'll fiddle with it for a bit and see if I can get it passing.

So it appears most of the remaining failures in test_kconv.rb are due to our NKF library still using Charset transcoding rather than jcodings. I'm going to resolve this and file another issue to get those fixed. That should give us the test coverage we want for this issue.