Index recycled string in str_dup() sep branch

NEWS.md

@@ -1,5 +1,8 @@
 # stringr (development version)
 
+* `str_dup()` now recycles `string` to the length of `times` when `sep` is
+  supplied, instead of failing with a subscript out of bounds error.
+
 # stringr 1.6.0
 
 ## Breaking changes

R/dup.R

@@ -22,7 +22,7 @@ str_dup <- function(string, times, sep = NULL) {
     out <- stri_dup(input$string, input$times)
   } else {
     out <- map_chr(seq_along(input$string), function(i) {
-      paste(rep(string[[i]], input$times[[i]]), collapse = sep)
+      paste(rep(input$string[[i]], input$times[[i]]), collapse = sep)
     })
   }
   names(out) <- names(input$string)

tests/testthat/_snaps/sub.md

@@ -4,10 +4,10 @@
       str_sub(x, 1:2, 1:3)
     Condition
       Error in `str_sub()`:
-      ! Can't recycle `string` (size 2) to match `end` (size 3).
+      ! Can't recycle `start` (size 2) to match `end` (size 3).
     Code
       str_sub(x, 1:2, 1:2) <- 1:3
     Condition
       Error in `str_sub<-`:
-      ! Can't recycle `string` (size 2) to match `value` (size 3).
+      ! Can't recycle `start` (size 2) to match `value` (size 3).
 

tests/testthat/test-dup.R

@@ -23,6 +23,11 @@ test_that("uses sep argument", {
 
   expect_equal(str_dup(character(), 1, sep = "-"), character())
   expect_equal(str_dup(character(), 2, sep = "-"), character())
+
+  expect_equal(
+    str_dup("na", 0:5, sep = "-"),
+    c("", "na", "na-na", "na-na-na", "na-na-na-na", "na-na-na-na-na")
+  )
 })
 
 test_that("separator must be a single string", {