On 03/09/2018 05:54, Ivan Gerasimov wrote:
Thanks Sherman and Stuart for the review!
On 9/2/18 2:45 PM, Stuart Marks wrote:
Yes, the fix itself looks fine. Quite subtle, good catch.
But should this have a regression test? I can imagine somebody coming
along later and "simplifying" (!(... > ...)) to (... <= ...) which
would reintroduce the bug.
Yes, there is a regression test added:
http://cr.openjdk.java.net/~igerasim/8210285/00/webrev/test/jdk/java/nio/charset/CharsetDecoder/NaNinCtor.java.html
The update to Charset-X-Coder.java.template looks okay. The changes also
CharsetEncoder so we'll need test coverage for that too.
-Alan
