jcheng5

Accepting POST requests from Shiny

(This post was motivated by a talk by @jnolis at CascadiaRConf 2021)

Recent versions of Shiny have an undocumented feature for handling POST requests that are not associated with any specific Shiny session. (Note that this functionality is missing our normal level of polish; it's a fairly low-level hook that I needed to make some things possible, but doesn't make anything easy.)

In a nutshell, it works by replacing your traditional ui object with a function(req), and then marking that function with an attribute indicating that it knows how to handle both GET and POST:

library(shiny)

jcheng5

assign("msg", local({
  i <- 0
  function(url, ...) {
    i <<- i + 1
    on.exit(i <<- i - 1)
    if (i > 1) {
      message("reentrancy detected!")
      return("reentrancy detected!")
    }
    cat("") # Calls R_ProcessEvents()

jcheng5

jpeg_dimensions <- function(filename, bytes = 1024) {
  bytes <- readBin(filename, "raw", n = bytes)
  
  if (length(bytes) < 2 || bytes[[1]] != 0xFF || bytes[[2]] != 0xD8) {
    stop("Couldn't decode jpeg")
  }
  
  ff <- which(bytes == 0xFF)
  c0 <- which(bytes == 0xC0)
  

jcheng5

```{js echo=FALSE}
function enshadowGtTable(tableEl) {
  var containerEl = tableEl.parentElement;
  var styleEl = containerEl.previousElementSibling;
  if (containerEl.tagName !== "DIV" || styleEl.tagName !== "STYLE") {
    throw new Error("Unexpected document structure");
  }
  var shadowDiv = document.createElement("div");
  shadowDiv.classList.add("gt-shadow-container");
  containerEl.parentElement.insertBefore(shadowDiv, styleEl);

jcheng5

As far as CRAN is concerned, there are two flavors of R on Solaris: one that is built using the Solaris Studio compiler, and one that is built using the GNU/gcc toolchain. The latter is far more up-to-date, but if your package requires it, then your DESCRIPTION file must declare that with the line SystemRequirements: GNU make.

These instructions are for configuring, building, and installing R-devel using the GNU/gcc toolchain (only).

You'll need VMWare Fusion on Mac, or VMWare Workstation (?) on Windows/Linux.

Get Solaris VM

Download the Solaris VM provided by Jeroen Ooms: https://github.com/jeroen/solarisvm

jcheng5

  590  pkgutil -i wget
  603  pkgutil -i gcc5gfortran
  609  pkgutil -i -y libiconv_dev
  613  pkgutil -i -y zlib_dev
  615  pkgutil -i -y libz_dev
  620  pkgutil -i -y liblzma_dev
  622  pkgutil -i -y libpcre_dev
  624  pkgutil -i -y libcurl_dev
  634  pkgutil -i -y libcurl4_dev
  647  pkgutil -y -i gmake

jcheng5

library(shiny)

ui <- fluidPage(
  numericInput("temp_c", "Celsius", NA),
  numericInput("temp_f", "Fahrenheit", NA)
)

server <- function(input, output, session) {
  c_to_f <- function(c, decimals = 1) {
    round((c * 9 / 5) + 32, decimals)

jcheng5

# BE SURE TO set your Mapbox access token with:
# options(mapbox.accessToken = "...")

library(htmltools)
library(leaflet)

mapboxgl_deps <- list(
  htmlDependency(
    "mapbox-gl-js", "0.53.0", c(href = "https://api.mapbox.com/mapbox-gl-js/v0.53.0/"),
    script = "mapbox-gl.js",

jcheng5

num_suffix <- function(x, base = 1000, suffixes = c("K", "M", "B", "T")) {
  
  if (length(suffixes) == 0) {
    tibble(
      scale_by = rep_len(1, length(x)),
      suffix = rep_len("", length(x))
    )
  }
  
  i <- floor(log(abs(x), base = base))

jcheng5

library(shiny)
library(leaflet)

#' Add a tile layer whose source is an R function
#'
#' @param tileFunc A function(x, y, z) that returns a 256x256 image object
#'   suitable for passing to `png::writePNG`.
#' @seealso [leaflet::addTiles()] for other parameters
addDynamicTiles <- function(map, tileFunc,
  layerId = paste0("leafletRaster", sample.int(9999999, 1)),