[R-pkg-devel] Fwd: Passing CRAN check when an examples must save a file

Jeff Newmiller jdnewm|| @end|ng |rom dcn@d@v|@@c@@u@
Sat Sep 28 07:41:42 CEST 2019


Read the Posting Guide... sending html formatted email corrupts your message.

Reading between the garbage...

a) Pasting an extension is not necessary.
b) using tempfile() in your example as an argument to sink( file=) is a straightforward way to illustrate that a file can be created.

On September 27, 2019 5:16:30 PM PDT, Tiago Olivoto <tiagoolivoto using gmail.com> wrote:
>Dear all,
>I had the same problem when I submitted my R package 'metan' to CRAN
>for
>the first time.
>
>"Please ensure that your functions do not write by default or in your
>examples/vignettes/tests in the user's home filespace (including the
>package directory and getwd()). In your examples/vignettes/tests you
>can
>write to tempdir()"
>
>I have several 'summary()' functions like:
>
>
>
>
>
>
>
>
>*summary.waasb <- function(object, export = FALSE, file.name
><http://file.name> = "summary_waasb", ...) {  # do all I need here  if
>(export == TRUE) {    sink(paste0(file.name <http://file.name>,
>".txt"))
># do all I need here    sink()  }}*
>
>All 'export' arguments are set to 'FALSE'. So, no file is generated in
>the
>user's working directory.
>Is the above function not according to the CRAN policies?
>
>In two vignettes I have a code like:
>
>
>
>*```{r, eval=FALSE }summary(stat_ge, export = TRUE)```*
>
>Please, note that 'eval = FALSE'. Could this be the problem?
>The suggestion is to write to a tempdir() but even looking at the
>Duncan
>Murdoch's answer I cannot see how to do that in my case.
>Looking for any suggestion.
>Best regards,
>Tiago Olivoto
>
>
>
>
>
>
>
>
>
>---------- Forwarded message ---------
>De: Duncan Murdoch <murdoch.duncan using gmail.com>
>Date: sex, 27 de set de 2019 às 09:44
>Subject: Re: [R-pkg-devel] Passing CRAN check when an examples must
>save a
>file
>To: Zachary Flynn <zlflynn using gmail.com>, <r-package-devel using r-project.org>
>
>
>On 26/09/2019 11:43 p.m., Zachary Flynn wrote:
>> I am working on a function in a package that saves a file.  I want to
>write
>> an example to document the function so it must save a file, but when
>I do,
>> the automatic CRAN checks reject the package because there is a
>> non-standard file saved in the directory when the second test runs
>(first,
>> it runs a test on Windows where the check error does not occur, then
>on
>> Linux where it does).  The relevant check output is:
>>
>> * checking for non-standard things in the check directory ... NOTE
>> Found the following files/directories:
>>    ‘Produc.csv’
>>
>> There is no file called Produc.csv anywhere in the directory (the
>> error doesn't show up in the first run on Windows).
>>
>> It is generated by running the example code which automatic CRAN
>> testing might be doing in between running the Windows and Linux
>> checks?
>>
>> How do I write an example that must save a file to be an example of
>> using the function without triggering this issue?
>
>Save the file to the temporary directory, ẹ.g.
>
>   filename <- file.path(tempdir(), "Prodục.csv")
>   write.csv(something, file = filename)
>
>You could clean up at the end of the example with
>
>   unlink(filename)
>
>but you shouldn't rely on this alone to prevent the error:  if you're
>writing files, you should *only* be doing it in the temporary
>directory.
>
>Duncan Murdoch
>
>______________________________________________
>R-package-devel using r-project.org mailing list
>https://stat.ethz.ch/mailman/listinfo/r-package-devel
>
>	[[alternative HTML version deleted]]
>
>______________________________________________
>R-package-devel using r-project.org mailing list
>https://stat.ethz.ch/mailman/listinfo/r-package-devel

-- 
Sent from my phone. Please excuse my brevity.



More information about the R-package-devel mailing list