r cookbook is defined as "r" everywhere, but here http://community.opscode.com/cookbooks/r it's an "R".

Bad luck, it breaks berkshelf. when it try to get r cookbook (with a "depend r" in metadata), I get:

`stat': No such file or directory - /tmp/d20130717-10057-1gmq8xf/r (Errno::ENOENT)

Because it's /tmp/d20130717-10057-1gmq8xf/R

What is your point of view? use "R" everywhere or change cookbook to "r" in opscode website (maybe with a remove / re-create).

Many thanks for this cookbook. Will it be uploaded to supermarket at some point?

Edit - Think I found it :-) https://supermarket.chef.io/cookbooks/r-project

Being able to install R packages via attributes would be nice:

node['r']['packages'] = ["Rserve", "RMySQL"]

I'd be happy to give it a try, but I'm not that familiar with chef yet.

Missing option to install packages from source URL. I got an issue while installing 'biom' package. It has been removed from the CRAN repository, so the r_package resource fails to install it.
Example suggestion:
r_package "biom" do
action :install
tarball_base_uri :'https://cran.r-project.org/src/contrib/Archive/biom/biom_0.3.12.tar.gz'
end
Thank you.

r_package won't halt convergence even when it fails to install a package. I submitted a fix in PR #36.

When I use r_package more than once, calls alternate between correctly and fully suppressing echoing from rinruby to allowing some echoing; something is being toggled on/off with each call.

I believe that libraries/default.rb should read

def r_package_installed?(package)
  require 'rinruby'
  r = RinRuby.new(:echo=>false)
  r.eval 'packages = installed.packages()[,1]'
  packages = r.pull 'packages'
  packages.include?(package)
end

See also this comment.

Installing via chef-solo(client?) means that a user can encounter the following:

install.packages(c('parallel'))
Installing package(s) into ‘/usr/local/lib/R/site-library’
(as ‘lib’ is unspecified)
Warning in install.packages(c("parallel")) :
  'lib = "/usr/local/lib/R/site-library"' is not writable
Would you like to use a personal library instead?  (y/n) n

This will likely require another recipe.
I am working on it.
Question is:
Which should be the default recipe?

• system install (as above)
• user install (pending pull request)

I tie down the version of all cookbooks used.
Using this cookbook makes it hard.
You are currently on version 0.3.0 metadata.rb but it is not tagged

Using the :upgrade action with the package provider in this cookbook results in excessive re-installation of packages because install_package is invoked unconditionally.

R will replace the installed package even if a newer version is unavailable, and this can cause problems for any jobs that attempt to make use of the package while it is being re-installed. Since Chef runs every half hour, this can happen frequently on a busy server.

I've written a short Ruby method to generate a bash command that returns true if the package needs to be updated. It does this by comparing the PACKAGES manifest against the DESCRIPTION of the package in the site-library.

def r_package_needs_update(package_name)
  available_packages_file = "/opt/R/src/contrib/PACKAGES"
  installed_package_manifest = "/usr/local/lib/R/site-library/#{package_name}/DESCRIPTION"
  sh_available_version = "awk '/^Package: #{package_name}/ {P=1} P==1 && /^Version:/ {print $2; exit}' #{available_packages_file}"
  sh_installed_version = "awk '/^Version/ {print $2}' #{installed_package_manifest}"
  "[ $(#{sh_available_version}) != $(#{sh_installed_version}) ]"
end

This seems to work, but would need to be generalized instead of hard-coding the paths. The check executes very quickly (~0.01s) compared to calling into R (~0.50s) but I'm not sure if this is a durable approach for checking package versions.

kubernetes/kubernetes#1662

When installing R with the source install_method i got the following error. I am using ubuntu 12.04.4

[2015-03-06T10:59:32-08:00] DEBUG: package[gcc-gfortran] checking package status for gcc-gfortran

* No version specified, and no candidate version available for gcc-gfortran
================================================================================
Error executing action `install` on resource 'package[gcc-gfortran]'
================================================================================

Chef::Exceptions::Package
-------------------------
No version specified, and no candidate version available for gcc-gfortran

Resource Declaration:
---------------------
# In /var/chef/cache/cookbooks/r/recipes/install_source.rb

 28: package "gcc-gfortran"
 29: 

Compiled Resource:
------------------
# Declared in /var/chef/cache/cookbooks/r/recipes/install_source.rb:28:in `from_file'

package("gcc-gfortran") do
  action :install
  retries 0
  retry_delay 2
  guard_interpreter :default
  package_name "gcc-gfortran"
  timeout 900
  cookbook_name "r"
  recipe_name "install_source"
end

[2015-03-06T10:59:33-08:00] INFO: Running queued delayed notifications before re-raising exception

The source installation method currently downloads the source tarball to /tmp.

/tmp may be mounted as a tmpfs, making it ill-suited for large files.

kubernetes/kubernetes#846

need support for repo installs of R for rhel based systems.

I set the version of R using the following attribute
node['r']['version'] = '3.2.0-1precise'

However, this gives me a converge error, the gist of it being:

The following packages have unmet dependencies:
r-base : Depends: r-recommended (= 3.2.0-1precise) but 3.2.0-4precise0 is to be installed
Recommends: r-base-html but it is not going to be installed
STDERR: E: Unable to correct problems, you have held broken packages.
---- End output of apt-get -q -y install r-base=3.2.0-1precise ----
Ran apt-get -q -y install r-base=3.2.0-1precise returned 100

I tried to install via the command line with the following command but got the same error

sudo apt-get install r-base=3.2.0-1precise

Installation succeeded with the following command instead:

sudo apt-get install r-base=3.2.0-1precise r-recommended=3.2.0-1precise

Hey @stevendanna,

I'm the author of the rstudio cookbook. I was wondering if it might make sense to either coordinate or combine our efforts. I'm currently working on an update that includes better APT support and a recipe for the Shiny server. I was going to fork and use the r_package LWRP or copy it, but instead have this set as a dependency. I think the only place our cookbooks overlap is the installation of r-base.

Just thought I'd say hello and get your thoughts. It might not make sense to combine them, but I think it makes sense to combine efforts on r-base and to ensure my usage of r_package is solid.

Cheers,

--Joe

All trivial points. Nonetheless:

Could the cookbook adopt a single name r? At the moment we have:

• r-project as the community site name.
• r as the github site name
• r-project as the recipe name
• R as the node key

Ruby convention suggests that
R_package should be r_package

Would you accept a pull request that made this uniform?
Would you accept r (assuming there is no cookbook named r that predates/has priority), or something else?
I am not aware of child dependencies, but this sort of change should justify a minor version bump.

I got an issue while installing 'biom' package. It has been removed from the CRAN repository, so the r_package resource fails to install it and does not report any kind of error.

For example: r_package 'biom'

This converges successfully , however the r package is not installed. The expected behaviour would be for this to fail as the package cannot be found.

Thank you.