Problem when using ASCII characters in R code

[maxl-amberg]

Hi,
I've got one question regaring one of my apps which runs perfectly fine when I use "run_dev.R" stored in the "dev" folder, but which always renders the following WARNING when I run devtools::check() in "03_deploy.R":
Found the following file with non-ASCII characters:
app_ui.R
mod_dataviz.R
Portable packages must use only ASCII characters in their R code,
except perhaps in comments.
Use \uxxxx escapes for other characters.

Since I would like to publish the app in German, my dataset contains variables that in turn contain Umlaute such as Ä, Ö, Ü, ä, ö, and ü (which all show up in both app_ui.R and mod_dataviz.R). @ColinFay : Despite having specified UTF-8 in the global options, do I have to fix the warning concerning these non-ASCII characters by replacing Ä, Ö, Ü, ä, ö, and ü in whatever way by \u00c4, \u00d6, \u00dc, \u00e4, \u00f6, and \u00fc respectively?

Note that replacing "RO-1: Ohne \nRückerstattung" by "RO-1: Ohne \nR"\u00fcckerstattung" in mod_dataviz.R solves the first warning regarding mod_dataviz.R but renders the following additional warning:
Found the following file with non-ASCII characters:
app_ui.R
Portable packages must use only ASCII characters in their R code,
except perhaps in comments.
Use \uxxxx escapes for other characters.
Warning in file 'R/mod_dataviz.R':
unable to translate 'RO-1: Ohne R<U+00FC>ckerstattung' to native encoding

Could replacing every Ä, Ö, Ü, ä, ö, and ü in app_ui.R and mod_dataviz.R work if instead done in the following fashion(?):
stringr::str_c("RO-1: Ohne \nR", "\u00fc", "ckerstattung") or
paste0("RO-1: Ohne \nR", "\u00fc", "ckerstattung")

Note that I actually do not want to sent the app/package to CRAN or to another package manager later on - I only want to deploy it e.g. via shinyapps.io or rstudioconnect. However, I do not know whether this warning is still a problem for this deploy-process.
Thanks a lot in advance!

[ColinFay]

Hey,

Yes, this is a known issue when it comes to publishing to the CRAN, which requires ASCII character only.
AFAIK, this is a legacy requirement, and almost every computer and server that would run this app will support non ASCII characters.

There are two solutions to this issue here:

• The one you pointed out, i.e. turning all non ASCII char to their ASCII equivalent. That can be done manually with

> stringi::stri_escape_unicode("Lélàô ")
[1] "L\\u00e9l\\u00e0\\u00f4 "

Note that the {prefixer} package can do that for you too https://github.com/dreamRs/prefixer/#not-ascii

• Ignore this warning ¯_(ツ)_/¯: if ever you only plan on releasing the app to a server and the server doesn't complain about the non ASCII, I suppose there is no harm in not escaping them all.

See also this chapter for more information https://r-pkgs.org/r.html#code-cran

Cheers,
Colin

[maxl-amberg]

Ah, I see.
Thanks a lot (also for mentioning additional resources)!

Best,
Maxl