Check shape operands for invalid dimension sizes

[burmako]

At the moment, RankedTensorType has the following verification logic (link):

LogicalResult
RankedTensorType::verify(function_ref<InFlightDiagnostic()> emitError,
                         ArrayRef<int64_t> shape, Type elementType,
                         Attribute encoding) {
  for (int64_t s : shape)
    if (s < 0 && !ShapedType::isDynamic(s))
      return emitError() << "invalid tensor dimension size";
  if (auto v = encoding.dyn_cast_or_null<VerifiableTensorEncoding>())
    if (failed(v.verifyEncoding(shape, elementType, emitError)))
      return failure();
  return checkTensorElementType(emitError, elementType);
}

It would be good to apply a similar level of validation to shape operands (like DynamicReshapeOp:: output_shape). If these operands are partially or fully constant, we could check that their elements are non-negative.