waterfall missing values are not respected

I would expect this to render to https://rreusser.github.io/plotly-mock-viewer/#waterfall_profit-loss_2018_positive-negative, but it currently doesn't

y <- c("Sales", "Consulting", "Maintenance", "Other revenue", "Net revenue", "Purchases", "Material expenses", "Personnel expenses", "Other expenses", "Operating profit", "Investment income", "Financial income", "Profit before tax", "Income tax (15%)", "Profit after tax")
  
  d <- data.frame(
    measure = c("relative", "relative", "relative", "relative", "total", "relative", "relative", "relative", "relative", "total", "relative", "relative", "total", "relative", "total"),
    y = factor(y, levels = y),
    x = c(375, 128, 78, 27, NA, -327, -12, -78, -12, NA, 32, 89, NA, -45, NA)
  )
  
plot_ly(d, measure = ~measure, y = ~y, x = ~x) %>%
    add_trace(type = "waterfall", orientation = "h") %>%
    layout(yaxis = list(autorange = "reversed"))

Reproduce the horizontal waterfall example from the issue using plot_ly, add_trace(type = "waterfall", orientation = "h"), and layout. Compare it with the expected plotly-mock-viewer result, then trace how the R binding passes NA values and waterfall measures. Done means the supplied example renders the missing values as expected.