Title: Systematically Run R CMD Checks
Version: 0.5.0
Description: Systematically Run R checks against multiple packages. Checks are run in parallel with strategies to minimize dependency installation. Provides out of the box interface for running reverse dependency check.
URL: https://Genentech.github.io/checked/, https://github.com/Genentech/checked
BugReports: https://github.com/Genentech/checked/issues
License: MIT + file LICENSE
Encoding: UTF-8
Depends: R (≥ 3.6.2)
Imports: callr, cli, glue, igraph, jsonlite, memoise, options, R6, rcmdcheck, rlang, utils (≥ 3.6.2), tools
RoxygenNote: 7.3.3
Suggests: remotes, testthat (≥ 3.0.0), visNetwork, withr
Config/Needs/website: r-lib/asciicast
Config/testthat/edition: 3
NeedsCompilation: no
Packaged: 2026-02-20 00:34:47 UTC; maksymis
Author: Szymon Maksymiuk ORCID iD [cre, aut], Doug Kelkhoff ORCID iD [aut], F. Hoffmann-La Roche AG [cph, fnd]
Maintainer: Szymon Maksymiuk <sz.maksymiuk@gmail.com>
Repository: CRAN
Date/Publication: 2026-02-20 10:00:17 UTC

checked: Systematically Run R CMD Checks

Description

Systematically Run R checks against multiple packages. Checks are run in parallel with strategies to minimize dependency installation. Provides out of the box interface for running reverse dependency check.

Author(s)

Maintainer: Szymon Maksymiuk sz.maksymiuk@gmail.com (ORCID)

Authors:

Other contributors:

See Also

Useful links:

Available packages database dependencies columns

Description

Available packages database dependencies columns

Usage

DB_COLNAMES

Format

An object of class character of length 6.

Dependencies categories

Description

Dependencies categories

Usage

DEP

Format

An object of class enum (inherits from factor) of length 5.

Strong dependencies categories

Description

Strong dependencies categories

Usage

DEP_STRONG

Format

An object of class enum (inherits from factor) of length 3.

Relation types

Description

A flag for edges that articulate different relations between nodes.

Usage

RELATION

Format

An object of class enum (inherits from factor) of length 2.

Check execution status categories

Description

Check execution status categories

Usage

STATUS

Format

An object of class enum (inherits from factor) of length 4.

Various utilities for formatting ANSI output

Description

Various utilities for formatting ANSI output

Usage

ansi_line_erase(n = "")

ansi_move_line_rel(n)

ansi_tty_height()

Arguments

n

The number of lines to move. Positive is up, negative is down.

Functions

A simple alternative to devtools::as.package

Description

Functionally identical to devtools' as.package, but without interactive options for package creation.

Usage

as.package.remotes(x)

Arguments

x

A package object to coerce

Details

Function required for communicating with the remotes package interface

Note

Code inspired by devtools load_pkg_description with very minor edits to further reduce devtools dependencies.

Convert a value to a set of dependency types

Description

Implements conventions established by both tools::package_dependencies() and pkgdepends::as_pkg_dependencies(), following pkgdepends::as_pkg_dependencies() return type structure of a list including ⁠$direct⁠ and ⁠$indirect⁠ dependency types. Reimplemented to avoid dependence on pkgdepends compilation requirements.

Usage

as_pkg_dependencies(x)

Arguments

x

A logical scalar, character string of "all", "most", "hard" or "soft", NA or a vector of dependency types.

Note

locally defined and bespoke dispatch system to avoid registering methods when loaded in combination with pkgdepends

Available Packages

Description

Functionally Equivalent to utils::available.packages(), assuming utils's cache doesn't expire in the middle of a top-level call evaluation. Modifications were made so that the results for queries with unique arguments are only called once for each top-level expression.

Usage

available_packages(...)

Details

Though utils::available.packages() will cache the PACKAGES index, it must still be parsed with each call. Since this can happen hundreds of times when building a ⁠R CMD check⁠ plan, this can cause a signficiant bottleneck to the responsiveness of this package. 

system.time({ for (i in 1:10) available.packages() })
#>    user  system elapsed
#>   3.453   0.196   3.655

system.time({ for (i in 1:10) available_packages() })
#>    user  system elapsed
#>   0.325   0.002   0.328

Note

This could be removed by propagating the available.packages() database matrix through all the calls that need to use it, though this would be a sizable refactor.

Check functions

Description

Set of functions to run orchestrated ⁠R CMD check⁠s and automatically manage the dependencies installation. Each functions prepares the plan based on the supplied package source(s) which includes installing dependencies and running required ⁠R CMD check⁠s. All the functions are parallelized through sperate processes

Arguments

path

file path to the package source directory

n

integer value indicating maximum number of subprocesses that can be simultaneously spawned when executing tasks.

output

character value specifying path where the output should be stored.

lib.loc

character vector with libraries allowed to be used when checking packages, defaults to entire .libPaths().

repos

character vector of repositories which will be used when generating task graph and later pulling dependencies.

reverse_repos

character vector of repositories which will be used to pull sources for reverse dependencies. In some cases, for instance using binaries on Linux, we want to use different repositories when pulling sources to check and different when installing dependencies.

restore

logical indicating whether output directory should be unlinked before running checks. If FALSE, an attempt will me made to restore previous progress from the same output

...

Additional arguments passed to run()

Value

checker() R6 class storing all the details regarding checks that run. Can be combined with results and summary() methods to generate results.

Check packages

Description

Runs classical ⁠R CMD check⁠ for the given source package. It first identifies and installs, in parallel, all dependencies required to check the package. Then, it runs ⁠R CMD check⁠ for each specified package.

Usage

check_pkgs(
  package,
  n = 2L,
  output = tempfile(paste(utils::packageName(), Sys.Date(), sep = "-")),
  lib.loc = .libPaths(),
  repos = getOption("repos"),
  restore = TRUE,
  ...
)

Arguments

package

A path to either package, directory with packages or name of the package (details)

n

integer value indicating maximum number of subprocesses that can be simultaneously spawned when executing tasks.

output

character value specifying path where the output should be stored.

lib.loc

character vector with libraries allowed to be used when checking packages, defaults to entire .libPaths().

repos

character vector of repositories which will be used when generating task graph and later pulling dependencies.

restore

logical indicating whether output directory should be unlinked before running checks. If FALSE, an attempt will me made to restore previous progress from the same output

...

Additional arguments passed to run()

Value

checker() R6 class storing all the details regarding checks that run. Can be combined with results and summary() methods to generate results.

See Also

Other checks: check_rev_deps(), checker, new_checker()

Check reverse dependencies

Description

Check a package's reverse dependencies in order to identify differences in reverse dependency check results when run alongside your package's development and release versions.

Usage

check_rev_deps(
  path,
  n = 2L,
  output = tempfile(paste(utils::packageName(), Sys.Date(), sep = "-")),
  lib.loc = .libPaths(),
  repos = getOption("repos"),
  reverse_repos = repos,
  restore = TRUE,
  ...
)

Arguments

path

file path to the package source directory

n

integer value indicating maximum number of subprocesses that can be simultaneously spawned when executing tasks.

output

character value specifying path where the output should be stored.

lib.loc

character vector with libraries allowed to be used when checking packages, defaults to entire .libPaths().

repos

character vector of repositories which will be used when generating task graph and later pulling dependencies.

reverse_repos

character vector of repositories which will be used to pull sources for reverse dependencies. In some cases, for instance using binaries on Linux, we want to use different repositories when pulling sources to check and different when installing dependencies.

restore

logical indicating whether output directory should be unlinked before running checks. If FALSE, an attempt will me made to restore previous progress from the same output

...

Additional arguments passed to run()

Details

Runs classical reverse dependency checks for the given source package. It first identifies reverse dependencies available in repos. Then, after installing all required dependencies, runs ⁠R CMD check⁠ twice for each package, one time with the release version of the given source package installed from repos and a second time with the development version installed from local source. Both ⁠R CMD checks⁠ are later compared to identify changes in reverse dependency behaviors.

Value

checker() R6 class storing all the details regarding checks that run. Can be combined with results and summary() methods to generate results.

See Also

Other checks: check_pkgs(), checker, new_checker()

Create a task to run ⁠R CMD check⁠

Description

Create a task to run ⁠R CMD check⁠

Usage

check_task(build_args = NULL, args = NULL, env = NULL, ...)

Arguments

build_args

Character vector of arguments to pass to ⁠R CMD build⁠. Pass each argument as a single element of this character vector (do not use spaces to delimit arguments like you would in the shell). For example, build_args = c("--force", "--keep-empty-dirs") is a correct usage and build_args = "--force --keep-empty-dirs" is incorrect.

args

Character vector of arguments to pass to ⁠R CMD check⁠. Pass each argument as a single element of this character vector (do not use spaces to delimit arguments like you would in the shell). For example, to skip running of examples and tests, use args = c("--no-examples", "--no-tests") and not args = "--no-examples --no-tests". (Note that instead of the --output option you should use the check_dir argument, because --output cannot deal with spaces and other special characters on Windows.)

env

A named character vector, extra environment variables to set in the check process.

...

Arguments passed on to task

.subclass

Additional subclasses.

See Also

Other tasks: install_task(), meta_task(), task()

R6 Checks Coordinator

Description

A stateful object that orchestrates all separate processes required to manage installation, library setup and run ⁠R CMD check⁠s in sequence.

Public fields

graph

(igraph::igraph())
A dependency graph, storing information about which dependencies are required prior to execution of each check task. Created with task_graph()

plan

(data.frame())
Checks task data.frame which is the source of all the checks.

output

(character(1))
Output directory where raw results and temporary library will be created and stored.

Methods

Public methods

Method new()

Initialize a new check design

Use checks data.frame to generate task graph in which all dependencies and installation order are embedded.

Usage
checker$new(
  plan,
  n = 2L,
  output = file.path(tempdir(), paste(packageName(), Sys.Date(), sep = "-")),
  lib.loc = .libPaths(),
  repos = getOption("repos"),
  restore = options::opt("restore"),
  ...
)
Arguments
plan

plan data.frame.

n

integer value indicating maximum number of subprocesses that can be simultaneously spawned when executing tasks.

output

character value specifying path where the output should be stored.

lib.loc

character vector with libraries allowed to be used when checking packages, defaults to entire .libPaths().

repos

character vector of repositories which will be used when generating task graph and later pulling dependencies.

restore

logical value, whether output directory should be unlinked before running checks. If FALSE, an attempt will me made to restore previous progress from the same output.

...

Additional arguments unused

Returns

checker.

Method active_processes()

Get Active Processes list

Usage
checker$active_processes()

Method failed_tasks()

Get Failed Tasks list

Usage
checker$failed_tasks()

Method terminate()

Kill All Active Design Processes

Immediately terminates all the active processes.

Usage
checker$terminate()

Method step()

Fill Available Processes with Tasks

Usage
checker$step()
Returns

A logical value, indicating whether processes are actively running.

Method start_next_task()

Start Next Task

Usage
checker$start_next_task()
Returns

A integer value, coercible to logical to indicate whether a new process was spawned, or -1 if all tasks have finished.

Method is_done()

Check if checks are done

Checks whether all the scheduled tasks were successfully executed.

Usage
checker$is_done()

Method clone()

The objects of this class are cloneable with this method.

Usage
checker$clone(deep = FALSE)
Arguments
deep

Whether to make a deep clone.

See Also

Other checks: check_pkgs(), check_rev_deps(), new_checker()

Examples

## Not run: 
library(checked)
plan <- plan_checks(c(
  system.file("example_packages", "exampleBad", package = "checked"),
  system.file("example_packages", "exampleGood", package = "checked")
))

orchestrator <- checker$new(
  plan,
  n = 10,
  repos = "https://cran.r-project.org/"
)

while (!orchestrator$is_done()) {
  orchestrator$start_next_task()
}

## End(Not run)

Parse R CMD checks from a partial check output string

Description

Parse R CMD checks from a partial check output string

Usage

checks_capture(x)

Arguments

x

A string, compsoed of any subsection of R CMD check console output

Value

A matrix of matches and capture groups "check" and "status" ("OK", "NONE", "NOTE", "WARNING" or "ERROR").

Examples

## Not run: 
check_output <- "
* checking check one ... OK
* checking check two ... NOTE
* checking tests ...
  Running test_abc.R
  Running test_xyz.R
 NONE
* checking check three ... WARNING
* ch
"

checks_capture(check_output)

## End(Not run)

Simplify Captures into Vector

Description

Simplify Captures into Vector

Usage

checks_simplify(x)

Arguments

x

Matrix of regex captures as produced by checks_capture.

Value

A vector of check status, with names indicating the check

Internal Utilities for Command-line Output

Description

Various helper functions for consistent cli output, including theming and formatting.

Usage

cli_table_row(
  status = "",
  ok = "OK",
  notes = "N",
  warnings = "W",
  errors = "E",
  msg = "",
  style = c("row", "title", "header"),
  symbols = DEFAULT_ROW_SYMBOL
)

cli_theme(..., .envir = parent.frame())

Arguments

status, ok, notes, warnings, errors

character[1L] A value to include in the respective columns of the table. Will be coerced to character if another type is provided.

msg

character[1L] A message to include to the right of the table row entry.

..., .envir

Additional arguments passed to cli::cli_div()

Reuse or Create A Null File Connection

Description

Reuse or Create A Null File Connection

Usage

devnull()

Produce cli output for a task

Description

Provided a task, allows for use of a handful of shorthand symbols which will use the task as a context for formatting task fields.

Usage

fmt(..., g, nodes, task = NULL, .envir = parent.frame(), ansi = TRUE)

Arguments

...

params passed to cli::format_inline.

g

task_graph object.

nodes

graph nodes to format.

task

task to format.

.envir

output environment.

ansi

logical whether ansi should be stripped.

Deduplicate attributes

Description

Primarily intended for cleaning up the result of an igraph::union(), which adds duplicated attributes when attributes of the same name exist in multiple graphs. Searches for suffixes and consolidates attributes, taking the attribute from the first non-NA value observed.

Usage

graph_dedup_attrs(g)

Arguments

g

task_graph object

Create a task to install a package and dependencies

Description

Create a task to install a package and dependencies

Usage

install_task(
  origin,
  type = package_install_type(origin),
  INSTALL_opts = NULL,
  lib = lib_path(origin),
  ...
)

Arguments

origin

pkg_origin() object.

type

character, indicating the type of package to download and install. Will be "source" except on Windows and some macOS builds: see the section on ‘Binary packages’ for those.

INSTALL_opts

an optional character vector of additional option(s) to be passed to R CMD INSTALL for a source package install. E.g., c("--html", "--no-multiarch", "--no-test-load") or, for macOS, "--dsym".

Can also be a named list of character vectors to be used as additional options, with names the respective package names.

lib

Any object that can be passed to lib() to generate a library path.

...

further arguments to be passed to download.file, available.packages, or to the functions for binary installs on macOS and Windows (which accept an argument "lock": see the section on ‘Locking’).

See Also

Other tasks: check_task(), meta_task(), task()

Get Library Location

Description

Get Library Location

Usage

lib(x, ...)

## S3 method for class ''NULL''
lib(x, ...)

## S3 method for class 'character'
lib(x, ...)

## S3 method for class 'lib_path_isolated'
lib(
  x,
  ...,
  lib.root = tempdir(),
  dir_hash = hash(Sys.time(), n = 8),
  name = ""
)

## S3 method for class 'lib_path_default'
lib(x, ..., lib.loc = .libPaths())

## S3 method for class 'task'
lib(x, ...)

## S3 method for class 'install_task'
lib(x, ...)

## S3 method for class 'check_task'
lib(x, ...)

Arguments

x

An object describing a library location

...

additional parameters passed to methods

lib.root

A root directory for the isolated library.

dir_hash

unique identifier of the isolated library

name

human-readable subname of the isolated library

lib.loc

Library paths, defaulting to .libPaths().

Make a Library Location

Description

A description of where packages should be installed. This object provides necessary information to determine where a package should be installed. lib_path method creates default path handlers for given pkg origin while lib_path_x creates an actual object.

Usage

lib_path(x, ..., .class = c())

lib_path_default(.class = c())

lib_path_isolated(.class = c())

Arguments

x

A pkg_origin() object used for default dispatch.

...

Additional values

.class

An optional subclass, used primarily for dispatch.

See Also

Other specs: pkg_origin()

Construct a 'Meta' Task

Description

Meta tasks are tasks which are not intended to perform computation. They exist simply to provide relationships among computational tasks.

Usage

meta_task(..., .subclass = NULL)

Arguments

...

Objects passed to specified class functions

.subclass

character name of the subclass. It will be appended with "_meta" suffix.

See Also

Other tasks: check_task(), install_task(), task()

Creating new Check Design Objects

Description

Instantiate a check design from a path or directory.

Usage

new_checker(...)

new_rev_dep_checker(x, ...)

Arguments

...

Additional arguments passed to new_checker()

x

A file path, passed to plan_rev_dep_checks()

See Also

Other checks: check_pkgs(), check_rev_deps(), checker

Other checks: check_pkgs(), check_rev_deps(), checker

checked Options

Description

Internally used, package-specific options. All options will prioritize R options() values, and fall back to environment variables if undefined. If neither the option nor the environment variable is set, a default value is used.

Checking Option Values

Option values specific to checked can be accessed by passing the package name to env. 

options::opts(env = "checked")

options::opt(x, default, env = "checked")

Options

tty_tick_interval

tty refresh interval when reporting results in milliseconds

default: 
0.1
option:

checked.tty_tick_interval

envvar:

R_CHECKED_TTY_TICK_INTERVAL (evaluated if possible, raw string otherwise)

tty_default_height

deafult tty height used for the ANSI reporter. Used only if correct values could not be acquired with system('tput lines')

default: 
50
option:

checked.tty_default_height

envvar:

R_CHECKED_TTY_DEFAULT_HEIGHT (evaluated if possible, raw string otherwise)

proactive_gc

logical, indicating whether additional garbage collection should be performed before starting a new task, if at least one process recently finalized. This can cause the checker to orchestrate tasks slower but is recommended to be used for designs with many sub-processes required as native garbage collection can lag leading to memory issues. Disable only when maximum prefromance is required and memory is not the issue.

default: 
TRUE
option:

checked.proactive_gc

envvar:

R_CHECKED_PROACTIVE_GC (evaluated if possible, raw string otherwise)

results_error_on

character vector indicating whether R error should be thrown when issues are discovered when generating results. "never" means that no errors are thrown. If "issues" then errors are emitted only on issues, whereas "potential issues" stands for error on both issues and potential issues.

default: 
"never"
option:

checked.results_error_on

envvar:

R_CHECKED_RESULTS_ERROR_ON (evaluated if possible, raw string otherwise)

results_keep

character vector indicating which packages should be included in the results. "all" means that all packages are kept. If "issues" then only packages with issues identified, whereas "potential_issues" stands for keeping packages with both "issues" and "potential_issues".

default: 
"all"
option:

checked.results_keep

envvar:

R_CHECKED_RESULTS_KEEP (evaluated if possible, raw string otherwise)

restore

logical indicating whether output directory should be unlinked before running checks. If FALSE, an attempt will me made to restore previous progress from the same output

default: 
NA
option:

checked.restore

envvar:

R_CHECKED_RESTORE (evaluated if possible, raw string otherwise)

add_remotes

logical indicating whether origins inheriting from pkg_origin_local, should be scanned for packages in the remotes field and added while constrocuting a plan task_grap

default: 
TRUE
option:

checked.add_remotes

envvar:

R_CHECKED_ADD_REMOTES (evaluated if possible, raw string otherwise)

check_envvars

named character vector of environment variables to use during the R CMD check.

default: 
c(`_R_CHECK_FORCE_SUGGESTS_` = "false", `_R_CHECK_RD_XREFS_` = "false", 
    `_R_CHECK_SYSTEM_CLOCK_` = "false", `_R_CHECK_SUGGESTS_ONLY_` = "true", 
    `_R_CHECK_CRAN_INCOMING_` = "false")
option:

checked.check_envvars

envvar:

R_CHECKED_CHECK_ENVVARS (evaluated if possible, raw string otherwise)

check_build_args

character vector of args passed to the R CMD build.

default: 
c("--no-build-vignettes", "--no-manual")
option:

checked.check_build_args

envvar:

R_CHECKED_CHECK_BUILD_ARGS (space-separated R CMD build flags)

check_args

character vector of args passed to the R CMD check.

default: 
c("--timings", "--ignore-vignettes", "--no-manual", "--as-cran")
option:

checked.check_args

envvar:

R_CHECKED_CHECK_ARGS (space-separated R CMD check flags)

See Also

options getOption Sys.setenv Sys.getenv

Other documentation: options_params

Checked Options

Description

Checked Options

Arguments

proactive_gc

logical, indicating whether additional garbage collection should be performed before starting a new task, if at least one process recently finalized. This can cause the checker to orchestrate tasks slower but is recommended to be used for designs with many sub-processes required as native garbage collection can lag leading to memory issues. Disable only when maximum prefromance is required and memory is not the issue. (Defaults to TRUE, overwritable using option 'checked.proactive_gc' or environment variable 'R_CHECKED_PROACTIVE_GC')

results_error_on

character vector indicating whether R error should be thrown when issues are discovered when generating results. "never" means that no errors are thrown. If "issues" then errors are emitted only on issues, whereas "potential issues" stands for error on both issues and potential issues. (Defaults to "never", overwritable using option 'checked.results_error_on' or environment variable 'R_CHECKED_RESULTS_ERROR_ON')

check_args

character vector of args passed to the R CMD check. (Defaults to c("--timings", "--ignore-vignettes", "--no-manual", "--as-cran"), overwritable using option 'checked.check_args' or environment variable 'R_CHECKED_CHECK_ARGS')

results_keep

character vector indicating which packages should be included in the results. "all" means that all packages are kept. If "issues" then only packages with issues identified, whereas "potential_issues" stands for keeping packages with both "issues" and "potential_issues". (Defaults to "all", overwritable using option 'checked.results_keep' or environment variable 'R_CHECKED_RESULTS_KEEP')

add_remotes

logical indicating whether origins inheriting from pkg_origin_local, should be scanned for packages in the remotes field and added while constrocuting a plan task_grap (Defaults to TRUE, overwritable using option 'checked.add_remotes' or environment variable 'R_CHECKED_ADD_REMOTES')

check_envvars

named character vector of environment variables to use during the R CMD check. (Defaults to ⁠c(⁠R_CHECK_FORCE_SUGGESTS⁠= "false",⁠R_CHECK_RD_XREFS⁠= "false", ; ⁠R_CHECK_SYSTEM_CLOCK⁠= "false",⁠R_CHECK_SUGGESTS_ONLY⁠= "true", ; ⁠R_CHECK_CRAN_INCOMING⁠ = "false")⁠, overwritable using option 'checked.check_envvars' or environment variable 'R_CHECKED_CHECK_ENVVARS')

tty_tick_interval

tty refresh interval when reporting results in milliseconds (Defaults to 0.1, overwritable using option 'checked.tty_tick_interval' or environment variable 'R_CHECKED_TTY_TICK_INTERVAL')

check_build_args

character vector of args passed to the R CMD build. (Defaults to c("--no-build-vignettes", "--no-manual"), overwritable using option 'checked.check_build_args' or environment variable 'R_CHECKED_CHECK_BUILD_ARGS')

restore

logical indicating whether output directory should be unlinked before running checks. If FALSE, an attempt will me made to restore previous progress from the same output (Defaults to NA, overwritable using option 'checked.restore' or environment variable 'R_CHECKED_RESTORE')

tty_default_height

deafult tty height used for the ANSI reporter. Used only if correct values could not be acquired with system('tput lines') (Defaults to 50, overwritable using option 'checked.tty_default_height' or environment variable 'R_CHECKED_TTY_DEFAULT_HEIGHT')

See Also

Other documentation: options()

Split a Filepath into Parts

Description

Split a Filepath into Parts

Usage

path_parts(x)

Arguments

x

A character(1L) or filepath

Value

A character vector of path parts

Build Package Dependencies Table

Description

Inspired by tools::package_dependencies, but with the added benefit of recording the dependency type and relationships throughout the dependency tree.

Usage

pkg_dependencies(
  packages,
  dependencies = TRUE,
  db = available_packages(),
  verbose = FALSE
)

Arguments

packages

a character vector of package names.

dependencies

A logical scalar, character string of "all", "most", "hard" or "soft", NA or a vector of dependency types compatible with as_pkg_dependencies() function.

db

character matrix as from available.packages() (with the default NULL the results of this call) or data frame variants thereof. Alternatively, a package database like the one available from https://cran.r-project.org/web/packages/packages.rds.

verbose

logical indicating if output should monitor the package search cycles.

Package specification

Description

Create package specification list which consists of all the details required to identify and acquire source of the package.

Usage

pkg_origin(package, ..., .class = c())

pkg_origin_repo(package, repos, ...)

pkg_origin_is_base(package, ...)

pkg_origin_base(package, ...)

pkg_origin_unknown(package, ...)

pkg_origin_local(path = NULL, ...)

pkg_origin_remote(remote = NULL, ...)

pkg_origin_archive(path = NULL, ...)

Arguments

package

name of the package.

...

parameters passed to downstream constructors.

.class

Additional subclasses.

repos

repository where package with given name should identified.

path

path to the source of the package (either bundled or not). URLs are acceptable.

remote

remote object from the remotes package used to identify non-standard packages.

See Also

Other specs: lib_path()

Plan R CMD Checks

Description

Generates a plan for running R CMD check for a specified set of packages.

Usage

plan_local_checks(package, repos = getOption("repos"))

Arguments

package

A path to either package, directory with packages or name of the package (details)

repos

repository used to identify packages when name is provided.

Details

package parameter has two different allowed values:

See Also

Other plan: plan_local_install(), plan_rev_dep_checks()

Plan source package installation

Description

Generates a plan for running installing a package from source.

Usage

plan_local_install(package, repos = getOption("repos"))

Arguments

package

A path to package source.

repos

repository used to identify packages when name is provided.

See Also

Other plan: plan_local_checks(), plan_rev_dep_checks()

Plan Reverse Dependency Checks

Description

Generates a plan for running reverse dependency check for certain source package. In such case path should be proivded with a directory path to the development version of the package and repos should be a repository for which reverse dependencies should be identified.

Usage

plan_rev_dep_checks(path, repos = getOption("repos"))

Arguments

path

path to the package source.

repos

repository used to identify reverse dependencies.

See Also

Other plan: plan_local_checks(), plan_local_install()

Print checked results

Description

Print checked results

Usage

## S3 method for class 'checked_results'
print(x, ...)

## S3 method for class 'rev_dep_dep_results'
print(x, ..., name = NULL, keep = options::opt("results_keep"))

## S3 method for class 'local_check_results'
print(x, ..., name = NULL, keep = options::opt("results_keep"))

Arguments

x

an object to be printed.

...

other parameters.

name

character name of the rev_dep_dep package

keep

character vector indicating which packages should be included in the results. "all" means that all packages are kept. If "issues" then only packages with issues identified, whereas "potential_issues" stands for keeping packages with both "issues" and "potential_issues". (Defaults to "all", overwritable using option 'checked.results_keep' or environment variable 'R_CHECKED_RESULTS_KEEP')

See Also

Other results: results()

Check checker Runner Reporters

Description

Reporters are used to configure how output is communicated while running a checker. They range from glossy command-line tools intended for displaying progress in an interactive R session, to line-feed logs which may be better suited for automated execution, such as in continuous itegration.

Usage

reporter_ansi_tty()

reporter_ansi_tty2()

reporter_basic_tty()

reporter_default()

Details

reporter_default()

Automatically chooses an appropriate reporter based on the calling context.

reporter_ansi_tty()

Highly dynamic output for fully capable terminals. Requires multi-line dynamic output, which may not be available in editors that that present a terminal as a web component.

reporter_basic_tty()

A line-feed reporter presenting output one line at a time, providing a reporter with minimal assumptions about terminal capabilities.

Reporter Internal Methods

Description

Each of the internal methods for reporters take a reporter, the check checker object and a calling environment.

Usage

report_sleep(reporter, checker, sleep)

## Default S3 method:
report_sleep(reporter, checker, sleep = 1)

report_start_setup(reporter, checker, ..., envir = parent.frame())

report_start_checks(reporter, checker, ..., envir = parent.frame())

## Default S3 method:
report_start_checks(reporter, checker, ..., envir = parent.frame())

report_status(reporter, checker, envir = parent.frame())

report_finalize(reporter, checker)

report_task(reporter, g, v)

report_step(reporter, checker)

Arguments

reporter

A object produced using reporters. Each reporter is a thin wrapper around an environment with a class name for dispatch. The reporter is mutable and captures any necessary state that needs to be tracked while reporting.

checker

checker The check checker to report as it evaluates.

sleep

numeric An interval to pause between reporter steps.

envir

environment An environment to attach to, to leverage on-exit hooks.

Check results

Description

Get R CMD check results

Usage

results(x, ...)

## S3 method for class 'checker'
results(x, error_on = options::opt("results_error_on"), ...)

## S3 method for class 'integer'
results(x, checker_obj, ...)

## S3 method for class 'igraph.vs'
results(x, ...)

## S3 method for class 'rev_dep_dep_meta_task'
results(x, checker_obj, ...)

## S3 method for class 'rev_dep_check_meta_task'
results(x, checker_obj, ...)

## S3 method for class 'local_check_meta_task'
results(x, checker_obj, ...)

Arguments

x

object which results should be presented.

...

other parameters.

error_on

character vector indicating whether R error should be thrown when issues are discovered when generating results. "never" means that no errors are thrown. If "issues" then errors are emitted only on issues, whereas "potential issues" stands for error on both issues and potential issues. (Defaults to "never", overwritable using option 'checked.results_error_on' or environment variable 'R_CHECKED_RESULTS_ERROR_ON')

checker_obj

checker object.

See Also

Other results: print.checked_results()

Run a Series of ⁠R CMD check⁠s

Description

run() provides a generic, and is the central interface for executing checkers. If a path is provided, a new reverse dependency check plan is generated from the source code path. Otherwise a plan can be built separately and executed using run().

Usage

run(checker, ..., reporter = reporter_default())

Arguments

checker

character or checker If a character value is provided, it is first coerced into a checker using new_rev_dep_checker().

...

Additional arguments passed to new_rev_dep_checker()

reporter

A reporter to provide progress updates. Will default to the most expressive command-line reporter given your terminal capabilities.

Create a 'cli' Spinner With Suppressed Output

Description

'cli' will implicitly push spinner output to various output streams, affecting the terminal cursor position. To allow for a terminal interface that has spinners above the last line, this function suppresses the output and simply returns its frame contents.

Usage

silent_spinner(..., stream = devnull())

Arguments

...

passed to cli::make_spinner

stream

passed to cli::make_spinner, defaults to a null file device

Value

A interface similar to a 'cli' spinner, but with suppressed output

Start a new task

Description

Starts task based on the task object encapsulated in the node taken from then task_graph g. It returns an install_process or check_process R6 object.

Usage

start_task(node, g, ...)

Arguments

node

Node(s) for which libpath should be constructed based on g

g

task_graph object

...

additional params passed to downstream methods

Task specification

Description

Create task specification list which consists of all the details required to run specific task.

Usage

task(..., .subclass = NULL)

Arguments

...

parameters passed to downstream constructors.

.subclass

Additional subclasses.

Details

Tasks can be nested, representing either a singular task, or a set of related tasks.

See Also

Other tasks: check_task(), install_task(), meta_task()

Task formatter bindings

Description

This bit of code is intended for use with fmt(), and allows for us to layer symbol bindings on top of the environment used for string interpolation which provide syntactic sugar for common formatting components.

Usage

task_formats(g = NULL, nodes = V(g), task = NULL, tasks = list(task))

Arguments

g

task_graph object.

nodes

graph nodes to format.

task

task to format.

tasks

currently unused.

Build task graph edges

Description

Edges describe relationships between tasks. Often, this is a dependency between packages, requiring that some package be installed before a latter task can be executed.

Usage

task_graph(x, repos = getOption("repos"), ...)

Arguments

x

a plan object, containing a list of related steps.

repos

repos, as expected by tools::package_dependencies() to determine package relationships.

...

params passed to helper methods.

Details

tools::package_dependencies() is used to calculate these relationships. However, the package data returned by utils::available.packages(), that is used internally to determine dependencies does not know about local or remote packages, so those are first appended to this data set prior to calculating edges. The bulk of this function serves to join this data.

Value

A data.frame that can be used to build igraph::make_graph edges.

Examples

## Not run: 
task_graph(plan_rev_dep_checks("."))

## End(Not run)

Libpaths from task graph

Description

Function that traverses over the task dependency task to acquire libpaths for given nodes. It ensures that when runing a node, a libpath is constructed which has all the required packages on it.

Usage

task_graph_libpaths(g, node = NULL, lib.loc = .libPaths(), output = tempdir())

Arguments

g

task_graph object

node

Node(s) for which libpath should be constructed based on g

lib.loc

Library paths, defaulting to .libPaths().

output

Path to the checked output directory

Find Task Neighborhood

Description

Find Task Neighborhood

Usage

task_graph_neighborhoods(g, nodes, ...)

Arguments

g

A task graph, as produced with task_graph()

nodes

Names or nodes objects of packages whose neighborhoods should be calculated.

Sort Task Graph by Strong Dependency Order

Description

Sort Task Graph by Strong Dependency Order

Usage

task_graph_sort(g)

Arguments

g

A igraph::graph, expected to contain node attribute type.

Value

The igraph::graph g, with vertices sorted in preferred installation order.

Note

Cyclic dependencies are possible. Cyclic dependencies are disallowed for all hard dependencies on CRAN today, though there have been historical instances where they appeared on CRAN.

Installation priority is based on:

  1. Total dependency footprint (low to high)

  2. Topology (leaf nodes first)

Find the Next Packages Not Dependent on an Unavailable Package

Description

While other packages are in progress, ensure that the next selected package already has its dependencies done.

Usage

task_graph_update_ready(
  g,
  v = V(g),
  dependencies = TRUE,
  status = STATUS$pending
)

Arguments

g

A dependency graph, as produced with task_graph().

v

Names or nodes objects of packages whose satisfiability should be checked.

dependencies

Which dependencies types should be met for a node to be considered satisfied.

status

status name. Nodes in v fill be filtered to consists only nodes with that status.

Details

There are helpers defined for particular use cases that strictly rely on the task_graph_which_ready(), they are:

Value

The name of the next package to prioritize

Find nodes with ready state

Description

List nodes which have ready state prioritizing check task nodes over install task nodes.

Usage

task_graph_which_ready(g)

Arguments

g

A dependency graph, as produced with task_graph().

Value

The names of packages with ready state.

Generate A Rate Limiting Throttle Function

Description

Generate A Rate Limiting Throttle Function

Usage

throttle(interval = 0.2)

Arguments

interval

An interval (in seconds) that is the minimum interval before throttle will return TRUE.

Value

A throttling function with the provided interval. When called, returns a logical value indicating whether the throttle interval has passed (TRUE if the interval has not yet passed).