[jdk21u-dev] RFR: 8357882: Use UTF-8 encoded data in LocaleDataTest

[Goetz Lindenmaier]

I backport this for parity with 21.0.10-oracle.

The change does not apply well. 
The cldr test data file underwent several changes in head.
Manual resolving is not a good approach here. 
Instead, I converted the file with the shell printf command. I verified it is identical 
in several places withoutb conflicts. I also looked at places with conflicts that they look good now.

The test also needed resolves due to context. 

Further, the test in 21 excercises two runs, one using a file for the COMPAT version that was removed in 25.
I adapted this as above, too.

I guess this is a true jewel for reviewers ... but keep in mind, it's only a test change :)

-------------

Commit messages:
 - backport 4fa4f15122213afea5cb25166c3b36a1c395b06c

Changes: https://git.openjdk.org/jdk21u-dev/pull/2371/files
  Webrev: https://webrevs.openjdk.org/?repo=jdk21u-dev&pr=2371&range=00
  Issue: https://bugs.openjdk.org/browse/JDK-8357882
  Stats: 8029 lines in 3 files changed: 0 ins; 101 del; 7928 mod
  Patch: https://git.openjdk.org/jdk21u-dev/pull/2371.diff
  Fetch: git fetch https://git.openjdk.org/jdk21u-dev.git pull/2371/head:pull/2371

PR: https://git.openjdk.org/jdk21u-dev/pull/2371