test: add SQL tests documenting Spark encode behavior

[andygrove]

Which issue does this PR close?

Closes #.

Rationale for this change

There is an upstream DataFusion PR to add a Spark-compatible encode expression: apache/datafusion#21331.

This PR adds tests to Comet to make it easier to review the DataFusion PR.

What changes are included in this PR?

Two files under spark/src/test/resources/sql-tests/expressions/string/:

• encode.sql runs on every supported Spark version. It covers UTF-8, US-ASCII, ISO-8859-1, UTF-16 (with BOM), UTF-16BE, UTF-16LE, and UTF-32 (no BOM), plus emoji / surrogate pairs, empty strings, NULL inputs for both arguments, case-insensitive charset names, column versus literal arguments, and binary input with both valid and invalid UTF-8 bytes.
• encode_strict.sql is gated by MinSparkVersion: 4.0. It pins Spark's charset whitelist (rejecting UTF-32BE, UTF-32LE, UTF8, UTF16, UTF16BE, ASCII, LATIN1, ISO88591, and EBCDIC with expect_error(INVALID_PARAMETER_VALUE.CHARSET)) and the raise-on-unmappable behavior (expect_error(MALFORMED_CHARACTER_CODING) for é in US-ASCII, Ā in ISO-8859-1, and an emoji in US-ASCII).

All positive queries use query spark_answer_only because Comet currently falls back to Spark for encode, and error cases use query expect_error(...) which works through the fallback path as well.

How are these changes tested?

Ran the new tests locally against both the default Spark 3.5 profile and the Spark 4.0 profile:

• ./mvnw test -Dsuites="org.apache.comet.CometSqlFileTestSuite encode" -Dtest=none passes encode.sql and skips encode_strict.sql (as expected, since the strict file is gated by MinSparkVersion: 4.0).
• ./mvnw -Pspark-4.0 test -Dsuites="org.apache.comet.CometSqlFileTestSuite encode" -Dtest=none passes both files.