category axis and shapes not playing well when data representing numbers and xref = "x"

Providing a list of strings representing numbers for category axis converts them to number rather than keeping them as strings.
This for me results in the inability to use shapes to shade part of the graph.

Example of what is not working:

library(plotly)

# Sample data
x <- c("1234", "2345", "3456", "4567")
y <- c(10, 15, 8, 12)

# Create a Plotly chart with a bar plot
fig <- plot_ly(x = ~x, y = ~y)

# Define a rectangle shape
shapes <- list(
  list(
    type = "rect",
    line = list(color = "red"),
    x0 = "1234", x1 = "2345",
    y0 = 0, y1 = 1,
    xref = "x",
    yref = "paper"
  )
)

# Add the rectangle shape to the layout
fig <- fig %>% layout(
  title = 'Highlighting with Lines', 
  xaxis = list(
    type = 'category',
    autotypenumbers = 'strict'
  ),
  shapes = shapes
)

# Display the chart
fig

Example of what is working:

library(plotly)

# Sample data
x <- c("#1234", "#2345", "#3456", "#4567")
y <- c(10, 15, 8, 12)

# Create a Plotly chart with a bar plot
fig <- plot_ly(x = ~x, y = ~y)

# Define a rectangle shape
shapes <- list(
  list(
    type = "rect",
    line = list(color = "red"),
    x0 = "#1234", x1 = "#2345",
    y0 = 0, y1 = 1,
    xref = "x",
    yref = "paper"
  )
)

# Add the rectangle shape to the layout
fig <- fig %>% layout(
  title = 'Highlighting with Lines', 
  xaxis = list(
    type = 'category',
    autotypenumbers = 'strict'
  ),
  shapes = shapes
)

# Display the chart
fig

Start by running the supplied R reproduction with plot_ly() and layout(), comparing numeric-looking category labels with the working labels prefixed by '#'. Trace how xaxis type='category' and autotypenumbers='strict' are handled for both bar data and shapes with xref='x'; done means numeric-looking strings remain categorical and the rectangle spans the intended categories.