Slow down successive exercise execution via debounce

NEWS.md

@@ -1,5 +1,6 @@
 # learnr (development version)
 
+-   Added option `tutorial.exercise.debounce` to slow down successive exercise execution (@internaut)
 -   Removed dependency on ellipsis (@olivroy, #809)
 -   Added Norwegian translation contributed by @jonovik. (#806)
 

R/exercise.R

@@ -33,9 +33,13 @@ setup_exercise_handler <- function(exercise_rx, session) {
   # setup reactive values for return
   rv <- reactiveValues(triggered = 0, result = NULL)
 
+  # debounce option to slow down successive exercise execution requests
+  debounce_ms <- getOption("tutorial.exercise.debounce", 0)   # value in milliseconds
+  if (debounce_ms > 0)
+    exercise_rx <- debounce(exercise_rx, debounce_ms)
+
   # observe input
   observeEvent(exercise_rx(), {
-
     # get exercise from app
     exercise <- exercise_rx()
     # Add tutorial information

pkgdown/assets/snippets/exercisedebounce.md

@@ -0,0 +1,2 @@
+# set debounce time to 1 second:
+options("tutorial.exercise.debounce" = 1000)

vignettes/articles/exercises.Rmd

@@ -80,6 +80,12 @@ This option can also be set either globally or per-chunk:
 insert_snippet("exerciseeval")
 ```
 
+It may be necessary to slow down successive exercise submissions, otherwise users may overwhelm your server by running exercise code chunks in fast repetition. You can do so globally by setting the option `tutorial.exercise.debounce` to an amount of time (in milliseconds) that is waited before the next code execution is performed:
+
+```{r snippet-exercisedebounce, echo = FALSE}
+insert_snippet("exercisedebounce")
+```
+
 ## Exercise Setup {#exercise-setup}
 
 Code chunks with `exercise=TRUE` are evaluated within standalone environments.