FULL PRODUCT VERSION :


A DESCRIPTION OF THE PROBLEM :
Nashorn version: hg tip 18edd7a1b166

The implementation for case-insensitive regular expressions does not comply to [ES5.1 - 15.10.2.8 Canonicalize]. This applies to both implementations, java.util.regex and Joni.

I've noticed this while working on integrating Joni for my project. You should be able to port the necessary changes from here [1]:
- UCS2Encoding.java contains the relevant changes for runtime.regexp.joni.EncodingHelper
- RegExpParser.java#L1256-L1412 contains the relevant changes for the java.util.regex based implementation

[1] https://github.com/anba/es6draft/tree/master/src/main/java/com/github/anba/es6draft/regexp

ADDITIONAL REGRESSION INFORMATION:
Java(TM) SE Runtime Environment (build 1.8.0-ea-b119)
Java HotSpot(TM) 64-Bit Server VM (build 25.0-b61, mixed mode)

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Test cases:

jjs> /[\u2160]/i.test("\u2170")

Expected: true
Actual: false


jjs> /\u0130/i.test("\u0069")

Expected: false
Actual: true


jjs> /\u1e9e/i.test("\u00df")

Expected: false
Actual: true


jjs> /\u0345{4}/i.test("\u0345\u0399\u03b9\u1fbe")

Expected: true
Actual: false


REPRODUCIBILITY :
This bug can be reproduced always.