Cope with ICU reporting obsolete language codes.

java.util.Locale mangles the new language codes for Hebrew and Indonesian into
the obsolete codes. I've added the ICU aliases so that ICU knows about the
obsolete codes (so that when we pass a mangled locale to ICU, ICU knows what
we're talking about). This patch adds tests, and also filters out the
duplicates caused by ICU now reporting both "he" and "iw", which we mangle
into "iw" and "iw". I've used a LinkedHashMap to preserve ICU's ordering;
there's no _need_ to do this, but it's nice behavior, better than the RI's.

Bug: 3452611
Change-Id: I6c6f3058902d5b7ee2f100605c019669da6893cd

0 files changed, 0 insertions, 0 deletions