Install Add-on Packages

Description

Utility for installing add-on packages.

Usage

R CMD INSTALL [options] [-l lib] pkgs

Arguments

Details

This will stop at the first error, so if you want all the pkgs to be tried, call this via a shell loop.

If used as R CMD INSTALL pkgs without explicitly specifying lib, packages are installed into the library tree rooted at the first directory in the library path which would be used by R run in the current environment.

To install into the library tree lib, use R CMD INSTALL -l lib pkgs. This prepends lib to the library path for duration of the install, so required packages in the installation directory will be found (and used in preference to those in other libraries).

Both lib and the elements of pkgs may be absolute or relative path names of directories. pkgs may also contain names of package archive files: these are then extracted to a temporary directory. These are tarballs containing a single directory, optionally compressed by gzip, bzip2, xz or compress. Finally, binary package archive files (as created by R CMD INSTALL --build) can be supplied.

Tarballs are by default unpackaged by the internal untar function: if needed an external tar command can be specified by the environment variable R_INSTALL_TAR: please ensure that it can handle the type of compression used on the tarball. (This is sometimes needed for tarballs containing invalid or unsupported sections, and can be faster on very large tarballs. Setting R_INSTALL_TAR to ‘⁠tar.exe⁠’ has been needed to overcome permissions issues on some Windows systems.)

The package sources can be cleaned up prior to installation by --preclean or after by --clean: cleaning is essential if the sources are to be used with more than one architecture or platform.

Some package sources contain a ‘configure’ script that can be passed arguments or variables via the option --configure-args and --configure-vars, respectively, if necessary. The latter is useful in particular if libraries or header files needed for the package are in non-system directories. In this case, one can use the configure variables LIBS and CPPFLAGS to specify these locations (and set these via --configure-vars), see section ‘Configuration variables’ in ‘R Installation and Administration’ for more information. (If these are used more than once on the command line they are concatenated.) The configure mechanism can be bypassed using the option --no-configure.

If the attempt to install the package fails, leftovers are removed. If the package was already installed, the old version is restored. This happens either if a command encounters an error or if the install is interrupted from the keyboard: after cleaning up the script terminates.

For details of the locking which is done, see the section ‘Locking’ in the help for install.packages.

Option --build can be used to tar up the installed package for distribution as a binary package (as used on macOS). This is done by utils::tar unless environment variable R_INSTALL_TAR is set.

By default a package is installed with static HTML help pages if and only if R was: use options --html and --no-html to override this.

Packages are not by default installed keeping the source formatting (see the keep.source argument to source): this can be enabled by the option --with-keep.source or by setting environment variable R_KEEP_PKG_SOURCE to yes.

Specifying the --install-tests option copies the contents of the ‘tests’ directory into the package installation. If the R_ALWAYS_INSTALL_TESTS environment variable is set to a true value, the tests will be installed even if --install-tests is omitted.

Use R CMD INSTALL --help for concise usage information, including all the available options.

Sub-architectures

An R installation can support more than one sub-architecture: currently this is most commonly used for 32- and 64-bit builds on Windows.

For such installations, the default behaviour is to try to install source packages for all installed sub-architectures unless the package has a configure script or a ‘src/Makefile’ (or ‘src/Makefile.win’ on Windows), when only compiled code for the sub-architecture running R CMD INSTALL is installed.

To install a source package with compiled code only for the sub-architecture used by R CMD INSTALL, use --no-multiarch. To install just the compiled code for another sub-architecture, use --libs-only.

There are two ways to install for all available sub-architectures. If the configure script is known to work for both Windows architectures, use flag --force-biarch (and packages can specify this via a ‘⁠Biarch: yes⁠’ field in their DESCRIPTION files). Second, a single tarball can be installed with

R CMD INSTALL --merge-multiarch mypkg_version.tar.gz

Staged installation

The default way to install source packages changed in R 3.6.0, so packages are first installed to a temporary location and then (if successful) moved to the destination library directory. Some older packages were written in ways that assume direct installation to the destination library.

Staged installation can currently be overridden by having a line ‘⁠StagedInstall: no⁠’ in the package's ‘DESCRIPTION’ file, via flag --no-staged-install or by setting environment variable R_INSTALL_STAGED to a false value (e.g. ‘⁠false⁠’ or ‘⁠no⁠’).

Staged installation requires either --pkglock or --lock, one of which is used by default.

Note

The options do not have to precede ‘⁠pkgs⁠’ on the command line, although it will be more legible if they do. All the options are processed before any packages, and where options have conflicting effects the last one will win.

Some parts of the operation of INSTALL depend on the R temporary directory (see tempdir, usually under ‘/tmp’) having both write and execution access to the account running R. This is usually the case, but if ‘/tmp’ has been mounted as noexec, environment variable TMPDIR may need to be set to a directory from which execution is allowed.

See Also

REMOVE; .libPaths for information on using several library trees; install.packages for R-level installation of packages; update.packages for automatic update of packages using the Internet or a local repository.

The chapter on ‘Add-on packages’ in ‘R Installation and Administration’ and the chapter on ‘Creating R packages’ in ‘Writing R Extensions’ via RShowDoc or in the ‘doc/manual’ subdirectory of the R source tree.