Add decimal remainder and checked_remainder for Spark arithmetic

[n0r0shi]

Description

Description:
Part of #3869

Description

Spark's % (remainder/modulo) operator supports decimal types, but Velox currently
only has remainder for integer and floating-point types — not for decimals.

In Spark's ANSI mode (spark.sql.ansi.enabled=true), decimal remainder should raise
an error on division by zero rather than returning null. This requires both a regular
and a checked variant.

This follows the same pattern established by:

• Add checked_add and checked_subtract for Spark decimal types #16301 / PR feat: Support decimal type for the Spark checked_add and checked_subtract functions #16302 — checked_add and checked_subtract for decimal
• Add checked_multiply for Spark decimal types #16306 / PR feat: Support decimal type for Spark checked_multiply function #16307 — checked_multiply for decimal
• Add checked_divide for Spark decimal types" #16322 / PR feat: Add checked_divide for Spark decimal arithmetic #16323 — checked_divide for decimal

Scope

• remainder for all decimal input combinations (short/long × short/long)
• checked_remainder for all decimal input combinations
• Both allowPrecisionLoss and denyPrecisionLoss variants
• Division by zero: null (non-checked) / throw (checked)
• int128 fast path with int256 fallback for rescaling overflow