[core] Use Locale.ROOT for machine-facing case conversions

[LuciferYang]

Purpose

close #9770

String.toLowerCase() and String.toUpperCase() follow the JVM default locale. Under a Turkish or Azeri default, i uppercases to the dotted İ and I lowercases to the dotless ı, so any token that is case-folded before being matched, parsed or signed stops matching what it is compared against. This is a real failure, not a theoretical one:

• CoreOptions.partitionMarkDoneActions() did PartitionMarkDoneAction.valueOf(x.replace('-','_').toUpperCase()), and both success-file and done-partition contain an i, so the default configuration threw IllegalArgumentException: No enum constant ...PartitionMarkDoneAction.SUCCESS_FİLE.
• RowKind.fromShortString("+i") uppercases to +İ, matches no case arm, and threw UnsupportedOperationException.
• DistributedLockDialectFactory lost SQLITE and MARIADB the same way, so JDBC catalog locking failed to resolve its dialect.
• OperatingSystem lowercases os.name and then looks for solaris, which becomes solarıs, so the OS came back UNKNOWN.
• HttpClient looks for the request-id header by lowercased name, so error messages silently lost the request id.
• System table lookup, StartupMode and Format enum parsing, and vector index type names are all exposed the same way, since manifests, INCREMENTAL, FROM_TIMESTAMP and IVF all contain an i.

The issue started from identifier matching: StringUtils.toLowerCaseIfNeed drives case-insensitive column and table matching, and CdcRecord.fieldNameLowerCase is the record side of that same join, so the two had to agree or a column silently nulled out.

Fixing three call sites and leaving the ones above would have been arbitrary, so this pins Locale.ROOT on every locale-sensitive case conversion in paimon-api, paimon-common, paimon-core and paimon-flink-cdc. Afterwards those modules have no bare String.toLowerCase() or toUpperCase() left.

BinaryString.toLowerCase and toUpperCase are untouched: their ASCII path uses Character.toLowerCase and the non-ASCII fallback already pins Locale.ROOT, so Paimon's data-level case conversion was already locale-independent.

Two changes that are not machine tokens, called out deliberately

StringUtils.toUpperCase / toLowerCase are what the CDC upper() and lower() computed columns run on, so this changes data written into the table for a tr, az or lt default locale: lower("ISTANBUL") now persists istanbul where it previously persisted ıstanbul. I think locale-independence is the behaviour you want there, since otherwise the value depends on which TaskManager ran the job, and it matches what BinaryString already does everywhere else. But it is a data change, not a token fix.

There is also an upgrade caveat for the narrow case of a table whose schema was inferred under such a locale. The persisted column name is ıd; after this change both sides of the join produce id, so the old column stops matching and schema evolution can append a second column beside it. TableNameConverter likewise resolves a different physical name. Being bug-compatible with a locale-dependent schema is not possible while also being correct, so this is a disclosure rather than something the patch works around.

Tests

TurkishLocaleParsingTest sets a Turkish default locale and covers the two failures reachable through a public entry point: parsing success-file,done-partition into the enum set, and RowKind.fromShortString("+i"). Only +i can discriminate, because Turkish differs from ROOT on i and I alone.

StringUtilsTest and CdcRecordTest cover the identifier path: toLowerCaseIfNeed("INDEX", false), toUpperCase("ıindex"), and the record-side key conversion.

FileFormat.getIdentifierPrefixOptions also stopped assuming that lowercasing preserves a string's length, which ROOT does not for İ; it now matches case-insensitively on the original key before slicing it.

Verified fail-on-base on JDK 11: with CoreOptions and RowKind reverted, the new test errors with No enum constant org.apache.paimon.CoreOptions.PartitionMarkDoneAction.SUCCESS_FİLE and Unsupported short string '+i' for row kind.. The four modules build clean with checkstyle and spotless enabled, and the ORC, Parquet and Avro format tests that consume the prefix options still pass.