[Rcpp-devel] What is /usr/bin/r ?

[Dirk Eddelbuettel]

Mike,

On 10 March 2014 at 14:41, Mike Willis wrote:
| 
| I'm making an rpm package to deploy to Linux machines I manage that 
| contains, amongst other things, a copy of the Rcpp package. The purpose 
| of this is that when a clsas that uses R is taught the lecturer doesn't 
| have to start by getting all the students to install all the required 
| packages. The package is made by installing all the packages myself, 
| then taking a copy of my R library and packaging that to put the files 
| in to the global R library directory. I've done this multiple times over 
| the years with no problem.
| 
| On this occasion one of the requested packages evidently drags in Rcpp 
| as a dependency and when I try to install my rpm package I get a 
| dependency error that nothing provides /usr/bin/r. I've traced this to 

It is from littler:

   edd at max:~$ dpkg -S /usr/bin/r
   littler: /usr/bin/r
   edd at max:~$ 

which, if memory serves, is also on some yum archives. I prefer Ubuntu and
Debian, and have it packaged there. 

But I also suspect your analysis may be wrong. See below:

| the Rcpp package which contains 44 files that begin
| 
| #!/usr/bin/r -t
| 
| Four randomly selected examples of  those 44 are:
| 
| examples/RcppInline/RcppInlineExample.r
| examples/FastLM/fastLMviaArmadillo.r

Example files are called by R during R CMD check, the shebang is merely a
courtesy for users who also want to call the installed file (and they still
need to make it chmod 0755)

| unitTests/runit.RObject.R
| unitTests/runit.Date.R

Ditto. 

All unit test files are called by the unit test framework (here: RUnit)
wrappers.

Just because you see a shebang should not imply a depend.

Can you share the exact error, please?
 
| Can someone tell me what /usr/bin/r refers to?
| 
| The executable for R is /usr/bin/R and it doesn't accept a -t option.

Operating systems for grown-ups are case-sensitive. /usr/bin/R != /usr/bin/r
 
| Please note I know next to nothing about R and even less about Rcpp.

Then maybe you should let someone else do the scripting, or at least ask for
local help. Adam Johansen and Yan Zhou over in Statistics are friends and may
be able to help you.

If this really was a hard constraint, CRAN would tell me in no uncertain
terms.  I believe you have a false positive, and should be able to tell rpm
to remain calm and carry on.

Lastly, if it really were a bug, we'd address.  So far I don't buy that it is
a bug.

Cheers, Dirk

-- 
Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com