Long type user input coercion missing implicit conversion from java.math.BigDecimal about sangria HOT 4 OPEN

  private case class TimestampCoercionViolation(value: String, typeName: String) extends ValueCoercionViolation(
    s"Positive Long value expected, but got $value of type $typeName"
  )

  implicit val TimestampType: ScalarType[Long] = ScalarType[Long](
    "Timestamp",
    description = Some(
      "The `Timestamp` scalar type represents non-fractional unsigned whole numeric values."
    ),
    coerceOutput = valueOutput,
    coerceUserInput = {
      case i: Int if i >= 0 => Right(i: Long)
      case i: Long if i >= 0L => Right(i)
      case i: BigInt if !i.isValidLong => Left(BigLongCoercionViolation)
      case i: BigInt if i >= 0 => Right(i.longValue)
      case d: Double if d.isWhole && d >= 0.0 => Right(d.toLong)
      case d: BigDecimal if d.isValidLong && d >= 0.0 => Right(d.longValue)
      case d: java.math.BigDecimal if BigDecimal(d).isValidLong && d.longValue >= 0.0 => Right(d.longValue)
      case x => Left(TimestampCoercionViolation(x.toString, x.getClass.toString))
    },
    coerceInput = {
      case ast.IntValue(i, _, _) => Right(i: Long)
      case ast.BigIntValue(i, _, _) if !i.isValidLong => Left(BigLongCoercionViolation)
      case ast.BigIntValue(i, _, _) => Right(i.longValue)
      case x => Left(TimestampCoercionViolation(x.toString, x.getClass.toString))
    }
  )

Positive Long value expected, but got 1680746052401 of type java.math.BigDecimal

  // There was an implicit to cut down on the wrapper noise for BigDec -> BigDecimal.
  // However, this may mask introduction of surprising behavior (e.g. lack of rounding
  // where one might expect it).  Wrappers should be applied explicitly with an
  // eye to correctness.