ICU-22919 Enable CI with JDK 21 and fix the json parsing

[mihnita]

At the code of the problem is a change in how Java parses strings to Double:

System.out.println(Double.valueOf("1234567890123456789"));

The results are:
jdk 17 : 1.23456789012345677E18
jdk 21 : 1.2345678901234568E18

The json parser was seeing a number and parsed it to a Double.
With the change the json parser sees a string and passes that on to our test suite, and the string is parsed / interpreted by our code.

So we just took control from the json parser.

Java supports a long value = 1234567890123456789L, there is no rounding, and the result from MessageFormatter is "1,234,567,890,123,456,789", as one would expect.

So the current fix does not need to change any code in MessageFormatter.

It only bypasses the json parsing of numbers and informs the test code on how to parse the string value.

Before (json parser in control):
value: 1234567890123456789
After (our test suite in control):
value: {"decimal": "1234567890123456789"}

Checklist

• Required: Issue filed: https://unicode-org.atlassian.net/browse/ICU-22919
• Required: The PR title must be prefixed with a JIRA Issue number.
• Required: The PR description must include the link to the Jira Issue, for example by completing the URL in the first checklist item
• Required: Each commit message must be prefixed with a JIRA Issue number.
• Issue accepted (done by Technical Committee after discussion)
• Tests included, if applicable
• API docs and/or User Guide docs changed or added, if applicable

[cjchapman]

I confirmed that all tests pass and I get BUILD SUCCESS when I run mvn verify locally with these changes. Thanks!

[mihnita]

I confirmed that all tests pass and I get BUILD SUCCESS when I run mvn verify locally with these changes. Thanks!

Thank you!
And thanks for finding it :-)