`mode_hdi()` returns an error

[ASKurz]

This holds for related functions, like hdi().

# Load
library(tidyverse)
library(ggdist)  # Version 3.3.3

# Following a grid approximation example from McElreath (2020, Section 3.2.2)

# How many grid points would you like?
n <- 1000
n_success <- 3
n_trials  <- 3

d <- tibble(p_grid = seq(from = 0, to = 1, length.out = n),
            prior  = 1) |> 
  mutate(likelihood = dbinom(n_success, size = n_trials, prob = p_grid)) |> 
  mutate(posterior  = (likelihood * prior) / sum(likelihood * prior))

# Sample from the grid
n_samples <- 1e4
set.seed(3)
samples <- d |> 
  slice_sample(n = n_samples, weight_by = posterior, replace = TRUE)

# Check the distribution in a plot
samples |> 
  ggplot(aes(x = p_grid)) +
  geom_density()

mode_hdi() returns an error for p_grid.

samples |> 
  mode_hdi(p_grid)

Error in quantile.default(dist_y, probs = 1 - .width, names = FALSE) : 
  missing values and NaN's not allowed if 'na.rm' is FALSE

What's the solution?

[ASKurz]

I don't know exactly when the change occurred, but this worked without an error as recently as ggdist 3.2.1.