[R-pkg-devel] file35451f8cc152 found a top level

Alexandre Génin @lex @ending from lec@irn@org
Wed Dec 19 12:00:33 CET 2018


On mardi 18 décembre 2018 22:19:52 CET Duncan Murdoch wrote:
> On 18/12/2018 4:08 PM, Alexandre Génin wrote:
> > Dear all,
> > 
> > I am preparing an update to my CRAN package 'spatialwarnings' and fixing
> > the last issues before submitting it.
> > 
> > I keep running into this type of message arising during checking:
> > 
> > * checking top-level files ... NOTE
> > 
> > Non-standard files/directories found at top level:
> >    'file35451f8cc152' 'file35451f9d8e7a' 'file3545214f4cdf'
> >    'file354524b354c4' 'file35452559ca7' 'file354525cb8eb1'
> >    'file35452a4e4bdd' 'file35452ac564a0' 'file35452ae421b2'
> >    'file35452d03ad' 'file35452dc8e4d4' 'file354530b2d171'
> > 
> > [snip]
> > 
> > (the complete log is available there: https://pastebin.com/JgyG9SwJ )
> > 
> > This happens for a variety of architectures, R versions and building
> > service (I tried rhub and win-builder). I am a bit puzzled because the
> > files are absent from the archive built through R CMD build and from my
> > git tree. I also added the following pattern to my .Rbuildignore to try
> > to fix the issue: ^file.*$, without success.
> > 
> > I have no idea of what generates these files (the problem does not arise
> > on my computers). Any hints on how to fix this ?
> > 
> > Thanks in advance and sorry if the question has already been asked (a
> > quick
> > search returned nothing)
> 
> Those look like filenames generated by the tempfile() function.  Those
> would normally be in the tempdir() directory.  If you have some
> environment variable (TMPDIR, TMP, TEMP) set to ".", they could be
> created in the top level of your package.  Alternatively, if you are
> using basename() on tempfile() you'll get the same sort of thing.
> 
> Duncan Murdoch

Ah, this was exactly it ! My checking script was using a temporary directory 
(using tempdir()) and those files were created in the package root for some 
reason. 

Thanks very much for the quick answer, 

Alexandre



More information about the R-package-devel mailing list