[SPARK-32624][SQL] Use getCanonicalName to fix byte[] compile issue
### What changes were proposed in this pull request?
```scala
scala> Array[Byte](1, 2).getClass.getName
res13: String = [B
scala> Array[Byte](1, 2).getClass.getCanonicalName
res14: String = byte[]
```
This pr replace `getClass.getName` with `getClass.getCanonicalName` in `CodegenContext.addReferenceObj` to fix `byte[]` compile issue:
```
...
/* 030 */ value_1 = org.apache.spark.sql.catalyst.util.TypeUtils.compareBinary(value_2, (([B) references[0] /* min */)) >= 0 && org.apache.spark.sql.catalyst.util.TypeUtils.compareBinary(value_2, (([B) references[1] /* max */)) <= 0;
/* 031 */ }
/* 032 */ return !isNull_1 && value_1;
/* 033 */ }
/* 034 */
/* 035 */
/* 036 */ }
20:49:54.886 WARN org.apache.spark.sql.catalyst.expressions.Predicate: Expr codegen error and falling back to interpreter mode
java.util.concurrent.ExecutionException: org.codehaus.commons.compiler.CompileException: File 'generated.java', Line 30, Column 81: failed to compile: org.codehaus.commons.compiler.CompileException: File 'generated.java', Line 30, Column 81: Unexpected token "[" in primary
...
```
### Why are the changes needed?
Fix compile issue when compiling generated code.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
Unit test.
Closes #29439 from wangyum/SPARK-32624.
Authored-by: Yuming Wang <yumwang@ebay.com>
Signed-off-by: Yuming Wang <wgyumg@gmail.com>
409fea30cc