This document will briefly cover the installation process of the header files for users since there are some possible environmental variables that can be used to aid in the download/installation of the CGAL header files on your machine. By default, and due to CRAN dependencies relying on the headers for C++ files, the package will download the latest header files from the CGAL GitHub. If you wish for more control or there is an error, read on!
The process of installing is, hopefully, quite simple for most users.
If you don’t have any strong desires, do the following:
install.packages("RcppCGAL")
cgal_install(cgal_path = NULL, version = NULL,
clean_files = TRUE, force = FALSE)The functions should take care of the rest! You will have the latest version of CGAL from GitHub. This should also happen by default on package installation if there is an internet connection.
If you have a particular version of the CGAL headers you would like
from the GitHub, use the version argument and keep
cgal_path = NULL.
install.packages("RcppCGAL")
cgal_install(cgal_path = NULL, version = "5.5.1",
clean_files = TRUE, force = FALSE)If you have already installed the CGAL header files and just want to
use that, you just need to specify the system path to the CGAL folder in
the argument cgal_path.
install.packages("RcppCGAL")
cgal_install(cgal_path = "path/to/CGAL", version = NULL,
clean_files = TRUE, force = FALSE)The function will then copy the files into the R package directory.
This is similar to the previous example; you just need to specify the exact location of the header files you want to download.
install.packages("RcppCGAL")
cgal_install(cgal_path = "https://some/url/cgalheaders.tar.gz", version = NULL,
clean_files = TRUE, force = FALSE)In some case, you may want to install the CGAL header files by default on installation and perhaps you may have a custom path. In this case, you can set two environmental variables.
The first, CGAL_DIR, will specify a directory or URL to
search for the CGAL header files like above. If it is not set, the
package will simply download the latest version or by using the argument
of cgal_path in the cgal_install function. In
the shell on Unix/Linux machines, you can do
export CGAL_DIR="https://some/url/cgalheaders.tar.gz"or you can set it in R directly
Sys.setenv(CGAL_DIR="https://some/url/cgalheaders.tar.gz")The second variable, CGAL_DOWNLOAD, will let R know to
install the header files automatically when installing the
RcppCGAL package. If CGAL_DOWNLOAD is any
other value than 0, the header files will be automatically downloaded.
If CGAL_DOWNLOAD=0 , the header files will not be
downloaded. This mean that in most cases, i.e. on CRAN, the header files
will be downloaded as long as there is an internet connection. You can
set this variable via the shell
or in R
export CGAL_DOWNLOAD="1"or you can set it in R directly
Sys.setenv(CGAL_DOWNLOAD="1")Then you can install the package and everything should flow automatically.
install.packages("RcppCGAL")The process of updating the files is simple. We can just follow the
same procedures as above using the cgal_install function
but set force = TRUE:
cgal_install(cgal_path = NULL, version = NULL,
clean_files = TRUE, force = TRUE)No need to install the whole package again. This will also give you a warning saying it will overwrite the current header files.
Pretty much you should always clean the header files—i.e., leave
clean_files = TRUE untouched. The reason is that the CGAL
header files have calls to std::exit,
std::cerr, and std::cout, which can can cause
R to crash or for messages to not appear on the R consol. By default,
RcppCGAL will search for these funtions and replace them
with the Rcpp equivalents; however, should you find this
isn’t happening, please raise an issue on the GitHub!