[hive] Fix Hive read failure for CHAR/VARCHAR columns exceeding Hive length limits

[thswlsqls]

Purpose

fix #9074

• PaimonObjectInspectorFactory.create() built PaimonCharObjectInspector/PaimonVarcharObjectInspector with the raw Paimon length, so a CHAR(>255) or VARCHAR(>65535) column made every Hive read of that table throw RuntimeException: ... out of allowed range.
• Clamp both to PaimonStringObjectInspector under the same conditions as the sibling HiveTypeUtils.PaimonToHiveTypeVisitor, so the inspector matches the string type HiveTypeUtils.toTypeInfo() already returns.
• That clamp came from [hive]Fix the issue of type conversion error from Paimon to Hive type #1568 (fix [Bug] Show create table failed in Hive when there is a field with type varchar(2147483646) in the table #1565) but was never applied to this factory, so the varchar(2147483646) case from [Bug] Show create table failed in Hive when there is a field with type varchar(2147483646) in the table #1565 still fails today.
• Lengths within the Hive limits are unchanged.

Tests

• Added PaimonObjectInspectorFactoryTest#testCreateCharVarcharObjectInspectorExceedingHiveLimit and #testCreateCharVarcharObjectInspectorWithinHiveLimit.
• mvn -pl paimon-hive/paimon-hive-connector-common clean install passed on JDK 11 — 102 unit tests and 83 *ITCase tests, plus checkstyle, spotless and enforcer.
• Reverting the fix and rerunning makes the three new positive assertions fail with the original exception.

[JingsongLi]

+1