API: Add tests for UnicodeUtil

[uros-b]

UnicodeUtil had no dedicated test class. This adds tests covering its public surface: isCharHighSurrogate at both ends of the high-surrogate range (0xD800-0xDBFF), plus a low surrogate and an ASCII char; truncateString when no truncation is needed, when truncating by code points, and that a surrogate pair (emoji) is not split; the IllegalArgumentException guard for a non-positive truncate length; truncateStringMin / truncateStringMax, including the overflow case where every retained code point is the maximum and no greater bound exists (returns null). Test-only changes.

[szehon-ho]

Looks good — test-only coverage for UnicodeUtil in the owning module, assertions match the implementation, and CI is green. Approving.

Optional follow-ups (not blocking): the String overload of truncateStringMax could also cover two paths that are already exercised for the Literal overload in TestMetricsTruncation:

1. No truncation needed (early return when the input already fits):

assertThat(UnicodeUtil.truncateStringMax("abc", 5)).isEqualTo("abc");

2. Carry when the last retained code point overflows:

String input =
    new StringBuilder()
        .append('a')
        .appendCodePoint(Character.MAX_CODE_POINT)
        .append('c')
        .toString();
assertThat(UnicodeUtil.truncateStringMax(input, 2)).isEqualTo("b");

[ebyhr]

I think there is one more uncovered path in UnicodeUtil.incrementCodePoint L127. When the last retained code point is \uD7FF (one below MIN_SURROGATE), the method skips over the surrogate block and returns \uE000.

  @Test
  void truncateStringMaxSkipsSurrogateRange() {
    // \uD7FF + 1 must skip the surrogate block (U+D800..U+DFFF) and return U+E000
    String input = "\uD7FF" + "extra";
    assertThat(UnicodeUtil.truncateStringMax(input, 1)).isEqualTo("\uE000");
  }

[uros-b]

Thank you folks @szehon-ho @ebyhr @nastra!