Migrate from IDNA2003 to UTS 45

[smola]

https://bugzilla.mozilla.org/show_bug.cgi?id=479520
https://codereview.chromium.org/23642003
http://blogs.msdn.com/b/shawnste/archive/2013/09/09/how-does-ie-handle-the-idn2008-rfcs.aspx

[smola]

This is not likely to be resolved any time soon. For now, we'll maintain the use of java.net.IDN, which implements IDNA2003.

[sideshowbarker]

This URL spec has now switched to referencing UTS 46 instead of RFC 3490 (and has dropped all references to RFC 3490 and replaced them with references to UTS 46). Specifically, it now references UTS 46 for ToASCII and ToUnicode.

I was going to say, "Time to switch away from java.net.IDN and on to com.ibm.icu.text.IDNA?" but as far as I can see, it seems that com.ibm.icu.text.IDNA also still just follows RFC 3490 for ToASCII and ToUnicode (instead of following UTS 46).

Note also the latest proposed update to UTS 46:

http://www.unicode.org/reports/tr46/proposed.html#ToASCII
http://www.unicode.org/reports/tr46/proposed.html#ToUnicode

[smola]

@sideshowbarker Thanks for the heads up.

http://icu-project.org/apiref/icu4j/com/ibm/icu/text/IDNA.html#getUTS46Instance(int)

ICU4J supports UTS 46 after all. I'll look into it. It's a pity to drop the "zero dependencies" claim, but it seems the only sane way to do it.

[sideshowbarker]

Good to hear that ICU4J supports UTS 46. Yeah it would be nice to have zero dependencies but I reckon a lot of target apps that'd use galimatias are probably already using ICU4J too. At least I can say that the validator is. So for me it wouldn't be introducing any new dependency