Unneccessary check for non-ASCII filenames should be removed

[toreeriksson]

Currently, boxr checks for non-ASCII filenames in the function box_filename() implemented in boxr__internal_misc.R, and raises an error if any non-ASCII characters are present. However, the BOX API does not have any such restrictions.

Thus, if lines 7-12 of boxr__internal_misc.R are deleted, the package will work for non-English environments as well. I have tested this locally and was able to upload a file named テストファイル.txt without problems.

diff --git i/R/boxr__internal_misc.R w/R/boxr__internal_misc.R
index 9e50756..2deb343 100644
--- i/R/boxr__internal_misc.R
+++ w/R/boxr__internal_misc.R
@@ -4,12 +4,6 @@
 # supported are those that contain non-printable ascii, / or \, names with
 # trailing spaces, and the special names "." and "..".
 box_filename <- function(x) {
-  x <- iconv(x, from = "", to = "ascii")
-
-  if (is.na(x))
-    stop("box.com accepts only valid ASCII filenames. Filename conversion to",
-         " ASCII via iconv() failed. See ?Encoding")
-
   if (nchar(x) > 255)
     stop("box.com accepts only filenames of 255 characters or less. Filename ",
          "is ", nchar(x), " characters long.")