Use revdep_check()
to run check_cran()
on all downstream
dependencies. Summarises the results with revdep_check_summary()
and
see problems with revdep_check_print_problems()
.
revdep_check_save_summary(pkg = ".")revdep_check_print_problems(pkg = ".")
revdep_check(pkg = ".", recursive = FALSE, ignore = NULL,
dependencies = c("Depends", "Imports", "Suggests", "LinkingTo"),
skip = character(), libpath = getOption("devtools.revdep.libpath"),
srcpath = libpath, bioconductor = FALSE, type = getOption("pkgType"),
threads = getOption("Ncpus", 1), env_vars = NULL, check_dir = NULL,
install_dir = NULL, quiet_check = TRUE)
revdep_check_resume(pkg = ".", ...)
revdep_check_reset(pkg = ".")
Path to package. Defaults to current directory.
If TRUE
look for full set of recursive dependencies.
A character vector of package names to ignore. These packages
will not appear in returned vector. This is used in
revdep_check
to avoid packages with installation problems
or extremely long check times.
A character vector listing the types of dependencies to follow.
A character vector of package names to exclude from the checks.
Path to library to store dependencies packages - if you you're doing this a lot it's a good idea to pick a directory and stick with it so you don't have to download all the packages every time.
Path to directory to store source versions of dependent packages - again, this saves a lot of time because you don't need to redownload the packages every time you run the package.
If TRUE
also look for dependencies amongst
bioconductor packages.
binary Package type to test (source, mac.binary etc). Defaults
to the same type as install.packages()
.
Number of concurrent threads to use for checking.
It defaults to the option "Ncpus"
or 1
if unset.
Environment variables set during R CMD check
A temporary directory to hold the results of the package checks. This should not exist as after the revdep checks complete successfully this directory is blown away.
Directory to store check and installation results.
If TRUE
, suppresses individual R CMD
check
output and only prints summaries. Set to FALSE
for
debugging.
Optionally, override original value of arguments to
revdep_check
. Use with care.
An invisible list of results. But you'll probably want to look
at the check results on disk, which are saved in check_dir
.
Summaries of all ERRORs and WARNINGs will be stored in
check_dir/00check-summary.txt
.
Install pkg
(in special library, see below).
Find all CRAN packages that depend on pkg
.
Install those packages, along with their dependencies.
Run R CMD check
on each package.
Uninstall pkg
(so other reverse dependency checks don't
use the development version instead of the CRAN version)
By default revdep_check
uses a temporary library to store any packages
that are required by the packages being tested. This ensures that they don't
interfere with your default library, but means that if you restart R
between checks, you'll need to reinstall all the packages. If you're
doing reverse dependency checks frequently, I recommend that you create
a directory for these packages and set options(devtools.revdep.libpath)
.
Revdep checks are resumable - this is very helpful if somethings goes
wrong (like you run out of power or you lose your internet connection) in
the middle of a check. You can resume a partially completed check with
revdep_check_resume()
, or blow away the cached result so you can
start afresh with revdep_check_reset()
.
revdep_maintainers()
to get a list of all revdep
maintainers.
# NOT RUN {
# Run R CMD check on all downstream dependencies
revdep_check()
revdep_check_save_summary()
revdep_check_print_problems()
# }
Run the code above in your browser using DataLab