-
Notifications
You must be signed in to change notification settings - Fork 2
Installation Information
Our packages are released to CRAN when new version are done. The GitHub README should link to the respective CRAN page. If you are interested in trying out development versions with very new features, read on.
We take great care in making sure that the GitHub hosted version are all buildable and include documented and working code. But of course, sometimes things are simply not done - hence the term "development version" instead of "official release". But if you notice anything out of order, please inform the respective maintainer, possibly best by adding a issue in the respective tracker. We don't bite and value constructive criticism.
Simply running
devtools::install_github("the_package_name", username="the_github_user_who_hosts_the_package")
will install the current GitHub version. The GitHub README of the package should include this line somewhere at the top of the page.
If the package contains C code and such stuff, Windows users must have RTools installed and properly setup, i.e., the necessary executables must be in the PATH.
PLEASE NOTE that we check that packages work together with their dependencies (that we control) by checking them with TravisCI. Here we check the versions on GitHub. Official packages and dependency chains are of course checked by CRAN. In summary, this means that for a certain GitHub package to work you might need to manually install its dependencies from our GitHub pages. Again, stick to official CRAN versions of all packages UNLESS you really have a valid reason to use a non-official release.
You can install a new package version after local code changes if you are in the checkout directory via
devtools::install(".")
Assuming you have a reasonably configured OS and R, you could also build and run tasks via the MAKEFILE. But only a VERY SMALL percentage of users should be interested in this.
-
Clone respective git repo from GitHub
-
Read this to get the Makefile tools: https://github.com/tudo-r/makeR/
-
Have recent version of R properly installed with all build tools. For Windows this will include RTools
-
Have R, Rscript (and the binaries of Rtools) in your PATH
-
Have roxygen2, devtools and testhat R packages installed
-
Possibly update all dependencies of the respective package that reside an GitHub as well. Install them via "make install".
-
In a console run "make install" to install. Done.
-
"make" will list all other build targets, e.g., targets for unit testing, checking, and so on.