cmake-modules - Man Page

Added in version 3.7.

Create a test that automatically loads specified data onto an Android device.

Use this module to push data needed for testing an Android device behavior onto a connected Android device. The module will accept files and libraries as well as separate destinations for each. It will create a test that loads the files into a device object store and link to them from the specified destination. The files are only uploaded if they are not already in the object store.

For example:

include(AndroidTestUtilities)
android_add_test_data(
  example_setup_test
  FILES <files>...
  LIBS <libs>...
  DEVICE_TEST_DIR "/data/local/tests/example"
  DEVICE_OBJECT_STORE "/sdcard/.ExternalData/SHA"
  )

At build time a test named "example_setup_test" will be created.  Run this test on the command line with ctest(1) to load the data onto the Android device.

android_add_test_data

android_add_test_data(<test-name>
  [FILES <files>...] [FILES_DEST <device-dir>]
  [LIBS <libs>...]   [LIBS_DEST <device-dir>]
  [DEVICE_OBJECT_STORE <device-dir>]
  [DEVICE_TEST_DIR <device-dir>]
  [NO_LINK_REGEX <strings>...]
  )

The android_add_test_data function is used to copy files and libraries needed to run project-specific tests. On the host operating system, this is done at build time. For on-device testing, the files are loaded onto the device by the manufactured test at run time.

This function accepts the following named parameters:

FILES <files>...

zero or more files needed for testing

LIBS <libs>...

zero or more libraries needed for testing

FILES_DEST <device-dir>

absolute path where the data files are expected to be

LIBS_DEST <device-dir>

absolute path where the libraries are expected to be

DEVICE_OBJECT_STORE <device-dir>

absolute path to the location where the data is stored on-device

DEVICE_TEST_DIR <device-dir>

absolute path to the root directory of the on-device test location

NO_LINK_REGEX <strings>...

list of regex strings matching the names of files that should be copied from the object store to the testing directory

Functions to help assemble a standalone bundle application.

A collection of CMake utility functions useful for dealing with .app bundles on the Mac and bundle-like directories on any OS.

The following functions are provided by this module:

fixup_bundle
copy_and_fixup_bundle
verify_app
get_bundle_main_executable
get_dotapp_dir
get_bundle_and_executable
get_bundle_all_executables
get_item_key
get_item_rpaths
clear_bundle_keys
set_bundle_key_values
get_bundle_keys
copy_resolved_item_into_bundle
copy_resolved_framework_into_bundle
fixup_bundle_item
verify_bundle_prerequisites
verify_bundle_symlinks

Requires CMake 2.6 or greater because it uses function, break and PARENT_SCOPE.  Also depends on GetPrerequisites.cmake.

DO NOT USE THESE Functions AT CONFIGURE TIME (from CMakeLists.txt)! Instead, invoke them from an install(CODE) or install(SCRIPT) rule.

fixup_bundle(<app> <libs> <dirs>)

Fix up <app> bundle in-place and make it standalone, such that it can be drag-n-drop copied to another machine and run on that machine as long as all of the system libraries are compatible.

If you pass plugins to fixup_bundle as the libs parameter, you should install them or copy them into the bundle before calling fixup_bundle. The <libs> parameter is a list of libraries that must be fixed up, but that cannot be determined by otool output analysis  (i.e. plugins).

Gather all the keys for all the executables and libraries in a bundle, and then, for each key, copy each prerequisite into the bundle.  Then fix each one up according to its own list of prerequisites.

Then clear all the keys and call verify_app on the final bundle to ensure that it is truly standalone.

Added in version 3.6: As an optional parameter (IGNORE_ITEM) a list of file names can be passed, which are then ignored (e.g. IGNORE_ITEM "vcredist_x86.exe;vcredist_x64.exe").

copy_and_fixup_bundle(<src> <dst> <libs> <dirs>)

Makes a copy of the bundle <src> at location <dst> and then fixes up the new copied bundle in-place at <dst>.

verify_app(<app>)

Verifies that an application <app> appears valid based on running analysis tools on it.  Calls message(FATAL_ERROR) if the application is not verified.

Added in version 3.6: As an optional parameter (IGNORE_ITEM) a list of file names can be passed, which are then ignored (e.g. IGNORE_ITEM "vcredist_x86.exe;vcredist_x64.exe")

get_bundle_main_executable(<bundle> <result_var>)

The result will be the full path name of the bundle's main executable file or an error: prefixed string if it could not be determined.

get_dotapp_dir(<exe> <dotapp_dir_var>)

Returns the nearest parent dir whose name ends with .app given the full path to an executable.  If there is no such parent dir, then simply return the dir containing the executable.

The returned directory may or may not exist.

get_bundle_and_executable(<app> <bundle_var> <executable_var> <valid_var>)

Takes either a .app directory name or the name of an executable nested inside a .app directory and returns the path to the .app directory in <bundle_var> and the path to its main executable in <executable_var>.

get_bundle_all_executables(<bundle> <exes_var>)

Scans <bundle> bundle recursively for all <exes_var> executable files and accumulates them into a variable.

get_item_key(<item> <key_var>)

Given <item> file name, generate <key_var> key that should be unique considering the set of libraries that need copying or fixing up to make a bundle standalone.  This is essentially the file name including extension with . replaced by _

This key is used as a prefix for CMake variables so that we can associate a set of variables with a given item based on its key.

clear_bundle_keys(<keys_var>)

Loop over the <keys_var> list of keys, clearing all the variables associated with each key.  After the loop, clear the list of keys itself.

Caller of get_bundle_keys should call clear_bundle_keys when done with list of keys.

set_bundle_key_values(<keys_var> <context> <item> <exepath> <dirs>
                      <copyflag> [<rpaths>])

Add <keys_var> key to the list (if necessary) for the given item. If added, also set all the variables associated with that key.

get_bundle_keys(<app> <libs> <dirs> <keys_var>)

Loop over all the executable and library files within <app> bundle (and given as extra <libs>) and accumulate a list of keys representing them.  Set values associated with each key such that we can loop over all of them and copy prerequisite libs into the bundle and then do appropriate install_name_tool fixups.

Added in version 3.6: As an optional parameter (IGNORE_ITEM) a list of file names can be passed, which are then ignored (e.g. IGNORE_ITEM "vcredist_x86.exe;vcredist_x64.exe")

copy_resolved_item_into_bundle(<resolved_item> <resolved_embedded_item>)

Copy a resolved item into the bundle if necessary. Copy is not necessary, if the <resolved_item> is "the same as" the <resolved_embedded_item>.

copy_resolved_framework_into_bundle(<resolved_item> <resolved_embedded_item>)

Copy a resolved framework into the bundle if necessary. Copy is not necessary, if the <resolved_item> is "the same as" the <resolved_embedded_item>.

By default, BU_COPY_FULL_FRAMEWORK_CONTENTS is not set.  If you want full frameworks embedded in your bundles, set BU_COPY_FULL_FRAMEWORK_CONTENTS to ON before calling fixup_bundle.  By default, COPY_RESOLVED_FRAMEWORK_INTO_BUNDLE copies the framework dylib itself plus the framework Resources directory.

fixup_bundle_item(<resolved_embedded_item> <exepath> <dirs>)

Get the direct/non-system prerequisites of the <resolved_embedded_item>. For each prerequisite, change the way it is referenced to the value of the _EMBEDDED_ITEM keyed variable for that prerequisite.  (Most likely changing to an @executable_path style reference.)

This function requires that the <resolved_embedded_item> be inside the bundle already.  In other words, if you pass plugins to fixup_bundle as the libs parameter, you should install them or copy them into the bundle before calling fixup_bundle.  The libs parameter is a list of libraries that must be fixed up, but that cannot be determined by otool output analysis.  (i.e., plugins)

Also, change the id of the item being fixed up to its own _EMBEDDED_ITEM value.

Accumulate changes in a local variable and make one call to install_name_tool at the end of the function with all the changes at once.

If the BU_CHMOD_BUNDLE_ITEMS variable is set then bundle items will be marked writable before install_name_tool tries to change them.

verify_bundle_prerequisites(<bundle> <result_var> <info_var>)

Verifies that the sum of all prerequisites of all files inside the bundle are contained within the bundle or are system libraries, presumed to exist everywhere.

Added in version 3.6: As an optional parameter (IGNORE_ITEM) a list of file names can be passed, which are then ignored (e.g. IGNORE_ITEM "vcredist_x86.exe;vcredist_x64.exe")

verify_bundle_symlinks(<bundle> <result_var> <info_var>)

Verifies that any symlinks found in the <bundle> bundle point to other files that are already also in the bundle...  Anything that points to an external file causes this function to fail the verification.

Check once whether the C compiler supports a given flag.

check_c_compiler_flag

check_c_compiler_flag(<flag> <resultVar>)

Check once that the <flag> is accepted by the compiler without a diagnostic. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

true indicates only that the compiler did not issue a diagnostic message when given the flag. Whether the flag has any effect is beyond the scope of this module.

Internally, try_compile() is used to perform the check. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the check compiles and links an executable program. If set to STATIC_LIBRARY, the check is compiled but not linked.

See also check_compiler_flag() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_c_compiler_flag(). Unknown flags in these variables can case a false negative result.

Added in version 3.19.

Check once whether the <lang> compiler supports a given flag.

check_compiler_flag

check_compiler_flag(<lang> <flag> <resultVar>)

Check once that the <flag> is accepted by the <lang> compiler without a diagnostic. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

true indicates only that the compiler did not issue a diagnostic message when given the flag. Whether the flag has any effect is beyond the scope of this module.

Internally, try_compile() is used to perform the check. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the check compiles and links an executable program. If set to STATIC_LIBRARY, the check is compiled but not linked.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_compiler_flag(). Unknown flags in these variables can case a false negative result.

Check once if C source code can be built.

check_c_source_compiles

check_c_source_compiles(<code> <resultVar>
                        [FAIL_REGEX <regex1> [<regex2>...]])

Check once that the source supplied in <code> can be built. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

If FAIL_REGEX is provided, then failure is determined by checking if anything in the compiler output matches any of the specified regular expressions.

Internally, try_compile() is used to compile the source. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the source is compiled and linked as an executable program. If set to STATIC_LIBRARY, the source is compiled but not linked. In any case, all functions must be declared as usual.

See also check_source_compiles() for a more general command syntax.

See also check_source_runs() to run compiled source.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_c_source_compiles():

Check once if given C source compiles and links into an executable and can subsequently be run.

check_c_source_runs

check_c_source_runs(<code> <resultVar>)

Check once that the source supplied in <code> can be built, linked as an executable, and then run. The <code> must contain at least a main() function.

The result is stored in the internal cache variable specified by <resultVar>. Success of build and run is indicated by boolean true. Failure to build or run is indicated by boolean false such as an empty string or an error message.

See also check_source_runs() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_c_source_runs():

Check once whether the CXX compiler supports a given flag.

check_cxx_compiler_flag

check_cxx_compiler_flag(<flag> <var>)

Check once that the <flag> is accepted by the compiler without a diagnostic. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

true indicates only that the compiler did not issue a diagnostic message when given the flag. Whether the flag has any effect is beyond the scope of this module.

Internally, try_compile() is used to perform the check. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the check compiles and links an executable program. If set to STATIC_LIBRARY, the check is compiled but not linked.

See also check_compiler_flag() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_cxx_compiler_flag(). Unknown flags in these variables can case a false negative result.

Check once if C++ source code can be built.

check_cxx_source_compiles

check_cxx_source_compiles(<code> <resultVar>
                          [FAIL_REGEX <regex1> [<regex2>...]])

Check once that the source supplied in <code> can be built. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

If FAIL_REGEX is provided, then failure is determined by checking if anything in the compiler output matches any of the specified regular expressions.

Internally, try_compile() is used to compile the source. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the source is compiled and linked as an executable program. If set to STATIC_LIBRARY, the source is compiled but not linked. In any case, all functions must be declared as usual.

See also check_source_compiles() for a more general command syntax.

See also check_source_runs() to run compiled source.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_cxx_source_compiles():

Check once if given C++ source compiles and links into an executable and can subsequently be run.

check_cxx_source_runs

check_cxx_source_runs(<code> <resultVar>)

Check once that the source supplied in <code> can be built, linked as an executable, and then run. The <code> must contain at least a main() function.

The result is stored in the internal cache variable specified by <resultVar>. Success of build and run is indicated by boolean true. Failure to build or run is indicated by boolean false such as an empty string or an error message.

See also check_source_runs() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_cxx_source_runs():

Check if a symbol exists as a function, variable, or macro in C++.

check_cxx_symbol_exists

check_cxx_symbol_exists(<symbol> <files> <variable>)

Check that the <symbol> is available after including given header <files> and store the result in a <variable>.  Specify the list of files in one argument as a semicolon-separated list. check_cxx_symbol_exists() can be used to check for symbols as seen by the C++ compiler, as opposed to check_symbol_exists(), which always uses the C compiler.

If the header files define the symbol as a macro it is considered available and assumed to work.  If the header files declare the symbol as a function or variable then the symbol must also be available for linking.  If the symbol is a type, enum value, or C++ template it will not be recognized: consider using the CheckTypeSize or CheckSourceCompiles module instead.

NOTE:

This command is unreliable when <symbol> is (potentially) an overloaded function. Since there is no reliable way to predict whether a given function in the system environment may be defined as an overloaded function or may be an overloaded function on other systems or will become so in the future, it is generally advised to use the CheckSourceCompiles module for checking any function symbol (unless somehow you surely know the checked function is not overloaded on other systems or will not be so in the future).

The following variables may be set before calling this macro to modify the way the check is run:

For example:

include(CheckCXXSymbolExists)

# Check for macro SEEK_SET
check_cxx_symbol_exists(SEEK_SET "cstdio" HAVE_SEEK_SET)
# Check for function std::fopen
check_cxx_symbol_exists(std::fopen "cstdio" HAVE_STD_FOPEN)

Added in version 3.3.

Check once whether the Fortran compiler supports a given flag.

check_fortran_compiler_flag

check_fortran_compiler_flag(<flag> <resultVar>)

Check once that the <flag> is accepted by the compiler without a diagnostic. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

true indicates only that the compiler did not issue a diagnostic message when given the flag. Whether the flag has any effect is beyond the scope of this module.

Internally, try_compile() is used to perform the check. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the check compiles and links an executable program. If set to STATIC_LIBRARY, the check is compiled but not linked.

See also check_compiler_flag() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_fortran_compiler_flag(). Unknown flags in these variables can case a false negative result.

Check if a Fortran function exists.

CHECK_FORTRAN_FUNCTION_EXISTS

CHECK_FORTRAN_FUNCTION_EXISTS(<function> <result>)

where

<function>

the name of the Fortran function

<result>

variable to store the result; will be created as an internal cache variable.

NOTE:

This command does not detect functions in Fortran modules. In general it is recommended to use CheckSourceCompiles instead to determine if a Fortran function or subroutine is available.

The following variables may be set before calling this macro to modify the way the check is run:

Added in version 3.1.

Check once if Fortran source code can be built.

check_fortran_source_compiles

check_fortran_source_compiles(<code> <resultVar>
    [FAIL_REGEX <regex>...]
    [SRC_EXT <extension>]
)

Check once that the source supplied in <code> can be built. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

If FAIL_REGEX is provided, then failure is determined by checking if anything in the compiler output matches any of the specified regular expressions.

By default, the test source file will be given a .F file extension. The SRC_EXT option can be used to override this with .<extension> instead-- .F90 is a typical choice.

See also check_source_compiles() for a more general command syntax.

See also check_source_runs() to run compiled source.

Internally, try_compile() is used to compile the source. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the source is compiled and linked as an executable program. If set to STATIC_LIBRARY, the source is compiled but not linked. In any case, all functions must be declared as usual.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_fortran_source_compiles():

Added in version 3.14.

Check once if given Fortran source compiles and links into an executable and can subsequently be run.

check_fortran_source_runs

check_fortran_source_runs(<code> <resultVar>
    [SRC_EXT <extension>])

Check once that the source supplied in <code> can be built, linked as an executable, and then run. The <code> must contain a Fortran program.

The result is stored in the internal cache variable specified by <resultVar>. Success of build and run is indicated by boolean true. Failure to build or run is indicated by boolean false such as an empty string or an error message.

check_fortran_source_runs("program test
real :: x[*]
call co_sum(x)
end program"
HAVE_COARRAY)

By default, the test source file will be given a .F90 file extension. The SRC_EXT option can be used to override this with .<extension> instead.

See also check_source_runs() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_fortran_source_runs():

Check once if a C function can be linked from system libraries.

check_function_exists

check_function_exists(<function> <variable>)

Checks that the <function> is provided by libraries on the system and store the result in internal cache variable <variable>.

The following variables may be set before calling this macro to modify the way the check is run:

NOTE:

Prefer using CheckSymbolExists or CheckSourceCompiles instead of this module, for the following reasons:

• check_function_exists() can't detect functions that are inlined in headers or specified as a macro.
• check_function_exists() can't detect anything in the 32-bit versions of the Win32 API, because of a mismatch in calling conventions.
• check_function_exists() only verifies linking, it does not verify that the function is declared in system headers.

Provides a macro to check if a header file can be included in CXX.

CHECK_INCLUDE_FILE_CXX

CHECK_INCLUDE_FILE_CXX(<include> <variable> [<flags>])

Check if the given <include> file may be included in a CXX source file and store the result in an internal cache entry named <variable>.  The optional third argument may be used to add compilation flags to the check (or use CMAKE_REQUIRED_FLAGS below).

The following variables may be set before calling this macro to modify the way the check is run:

See modules CheckIncludeFile and CheckIncludeFiles to check for one or more C headers.

Provides a macro to check if a header file can be included in C.

CHECK_INCLUDE_FILE

CHECK_INCLUDE_FILE(<include> <variable> [<flags>])

Check if the given <include> file may be included in a C source file and store the result in an internal cache entry named <variable>.  The optional third argument may be used to add compilation flags to the check (or use CMAKE_REQUIRED_FLAGS below).

The following variables may be set before calling this macro to modify the way the check is run:

See the CheckIncludeFiles module to check for multiple headers at once.  See the CheckIncludeFileCXX module to check for headers using the CXX language.

Provides a macro to check if a list of one or more header files can be included together.

CHECK_INCLUDE_FILES

CHECK_INCLUDE_FILES("<includes>" <variable> [LANGUAGE <language>])

Check if the given <includes> list may be included together in a source file and store the result in an internal cache entry named <variable>.  Specify the <includes> argument as a ;-list of header file names.

If LANGUAGE is set, the specified compiler will be used to perform the check. Acceptable values are C and CXX. If not set, the C compiler will be used if enabled. If the C compiler is not enabled, the C++ compiler will be used if enabled.

The following variables may be set before calling this macro to modify the way the check is run:

See modules CheckIncludeFile and CheckIncludeFileCXX to check for a single header file in C or CXX languages.

Added in version 3.9.

Check whether the compiler supports an interprocedural optimization (IPO/LTO). Use this before enabling the INTERPROCEDURAL_OPTIMIZATION target property.

check_ipo_supported

check_ipo_supported([RESULT <result>] [OUTPUT <output>]
                    [LANGUAGES <lang>...])

Options are:

RESULT <result>

Set <result> variable to YES if IPO is supported by the compiler and NO otherwise.  If this option is not given then the command will issue a fatal error if IPO is not supported.

OUTPUT <output>

Set <output> variable with details about any error.

LANGUAGES <lang>...

Specify languages whose compilers to check. Languages C, CXX, and Fortran are supported.

It makes no sense to use this module when CMP0069 is set to OLD so module will return error in this case. See policy CMP0069 for details.

Added in version 3.13: Add support for Visual Studio generators.

Added in version 3.24: The check uses the caller's CMAKE_<LANG>_FLAGS and CMAKE_<LANG>_FLAGS_<CONFIG> values. See policy CMP0138.

check_ipo_supported() # fatal error if IPO is not supported
set_property(TARGET foo PROPERTY INTERPROCEDURAL_OPTIMIZATION TRUE)

# Optional IPO. Do not use IPO if it's not supported by compiler.
check_ipo_supported(RESULT result OUTPUT output)
if(result)
  set_property(TARGET foo PROPERTY INTERPROCEDURAL_OPTIMIZATION TRUE)
else()
  message(WARNING "IPO is not supported: ${output}")
endif()

Check whether a language can be enabled by the enable_language() or project() commands:

check_language

check_language(<lang>)

Try enabling language <lang> in a test project and record results in the cache:

CMAKE_<LANG>_COMPILER

If the language can be enabled, this variable is set to the compiler that was found.  If the language cannot be enabled, this variable is set to NOTFOUND.

If this variable is already set, either explicitly or cached by a previous call, the check is skipped.

CMAKE_<LANG>_HOST_COMPILER

This variable is set when <lang> is CUDA or HIP.

If the check detects an explicit host compiler that is required for compilation, this variable will be set to that compiler. If the check detects that no explicit host compiler is needed, this variable will be cleared.

If this variable is already set, its value is preserved only if CMAKE_<LANG>_COMPILER is also set. Otherwise, the check runs and overwrites CMAKE_<LANG>_HOST_COMPILER with a new result. Note that CMAKE_<LANG>_HOST_COMPILER documents it should not be set without also setting CMAKE_<LANG>_COMPILER to a NVCC compiler.

CMAKE_<LANG>_PLATFORM

This variable is set to the detected GPU platform when <lang> is HIP.

If the variable is already set its value is always preserved. Only compatible values will be considered for CMAKE_<LANG>_COMPILER.

For example:

check_language(Fortran)
if(CMAKE_Fortran_COMPILER)
  enable_language(Fortran)
else()
  message(STATUS "No Fortran support")
endif()

Check once if the function exists in system or specified library.

CHECK_LIBRARY_EXISTS

CHECK_LIBRARY_EXISTS(LIBRARY FUNCTION LOCATION VARIABLE)

LIBRARY  - the name of the library you are looking for
FUNCTION - the name of the function
LOCATION - location where the library should be found
VARIABLE - internal cache variable to store the result

Prefer using CheckSymbolExists or CheckSourceCompiles instead of this module for more robust detection if a function is available in a library.

The following variables may be set before calling this macro to modify the way the check is run:

Added in version 3.18.

Check whether the compiler supports a given link flag.

check_linker_flag

check_linker_flag(<lang> <flag> <var>)

Check that the link <flag> is accepted by the <lang> compiler without a diagnostic.  Stores the result in an internal cache entry named <var>.

This command temporarily sets the CMAKE_REQUIRED_LINK_OPTIONS variable and calls the check_source_compiles() command from the CheckSourceCompiles module.  See that module's documentation for a listing of variables that can otherwise modify the build.

The underlying implementation relies on the LINK_OPTIONS property to check the specified flag. The LINKER: prefix, as described in the target_link_options() command, can be used as well.

A positive result from this check indicates only that the compiler did not issue a diagnostic message when given the link flag.  Whether the flag has any effect or even a specific one is beyond the scope of this module.

NOTE:

Added in version 3.16.

Check once whether the Objective-C compiler supports a given flag.

check_objc_compiler_flag

check_objc_compiler_flag(<flag> <resultVar>)

Check once that the <flag> is accepted by the compiler without a diagnostic. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

true indicates only that the compiler did not issue a diagnostic message when given the flag. Whether the flag has any effect is beyond the scope of this module.

Internally, try_compile() is used to perform the check. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the check compiles and links an executable program. If set to STATIC_LIBRARY, the check is compiled but not linked.

See also check_compiler_flag() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_objc_compiler_flag(). Unknown flags in these variables can case a false negative result.

Added in version 3.16.

Check once if Objective-C source can be built.

check_objc_source_compiles

check_objc_source_compiles(<code> <resultVar>
                           [FAIL_REGEX <regex1> [<regex2>...]])

Check once that the source supplied in <code> can be built. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

If FAIL_REGEX is provided, then failure is determined by checking if anything in the compiler output matches any of the specified regular expressions.

Internally, try_compile() is used to compile the source. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the source is compiled and linked as an executable program. If set to STATIC_LIBRARY, the source is compiled but not linked. In any case, all functions must be declared as usual.

See also check_source_compiles() for a more general command syntax.

See also check_source_runs() to run compiled source.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_objc_source_compiles()

Added in version 3.16.

Check once if given Objective-C source compiles and links into an executable and can subsequently be run.

check_objc_source_runs

check_objc_source_runs(<code> <resultVar>)

Check once that the source supplied in <code> can be built, linked as an executable, and then run. The <code> must contain at least a main() function.

The result is stored in the internal cache variable specified by <resultVar>. Success of build and run is indicated by boolean true. Failure to build or run is indicated by boolean false such as an empty string or an error message.

See also check_source_runs() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_objc_source_runs()

Added in version 3.16.

Check once whether the Objective-C++ compiler supports a given flag.

check_objcxx_compiler_flag

check_objcxx_compiler_flag(<flag> <resultVar>)

Check once that the <flag> is accepted by the compiler without a diagnostic. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

true indicates only that the compiler did not issue a diagnostic message when given the flag. Whether the flag has any effect is beyond the scope of this module.

Internally, try_compile() is used to perform the check. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the check compiles and links an executable program. If set to STATIC_LIBRARY, the check is compiled but not linked.

See also check_compiler_flag() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_objcxx_compiler_flag(). Unknown flags in these variables can case a false negative result.

Added in version 3.16.

Check once if Objective-C++ source can be built.

check_objcxx_source_compiles

check_objcxx_source_compiles(<code> <resultVar>
                             [FAIL_REGEX <regex1> [<regex2>...]])

Check once that the source supplied in <code> can be built. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

If FAIL_REGEX is provided, then failure is determined by checking if anything in the compiler output matches any of the specified regular expressions.

Internally, try_compile() is used to compile the source. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the source is compiled and linked as an executable program. If set to STATIC_LIBRARY, the source is compiled but not linked. In any case, all functions must be declared as usual.

See also check_source_compiles() for a more general command syntax.

See also check_source_runs() to run compiled source.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_objcxx_source_compiles()

Added in version 3.16.

Check once if given Objective-C++ source compiles and links into an executable and can subsequently be run.

check_objcxx_source_runs

check_objcxx_source_runs(<code> <resultVar>)

Check once that the source supplied in <code> can be built, linked as an executable, and then run. The <code> must contain at least a main() function.

The result is stored in the internal cache variable specified by <resultVar>. Success of build and run is indicated by boolean true. Failure to build or run is indicated by boolean false such as an empty string or an error message.

See also check_source_runs() for a more general command syntax.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_objcxx_source_runs()

Added in version 3.14.

Check whether the linker supports Position Independent Code (PIE) or No Position Independent Code (NO_PIE) for executables. Use this to ensure that the POSITION_INDEPENDENT_CODE target property for executables will be honored at link time.

check_pie_supported

check_pie_supported([OUTPUT_VARIABLE <output>]
                    [LANGUAGES <lang>...])

Options are:

OUTPUT_VARIABLE <output>

Set <output> variable with details about any error. If the check is bypassed because it uses cached results from a previous call, the output will be empty even if errors were present in the previous call.

LANGUAGES <lang>...

Check the linkers used for each of the specified languages. If this option is not provided, the command checks all enabled languages.

C, CXX, Fortran are supported.

Added in version 3.23: OBJC, OBJCXX, CUDA, and HIP are supported.

It makes no sense to use this module when CMP0083 is set to OLD, so the command will return an error in this case.  See policy CMP0083 for details.

For each language checked, two boolean cache variables are defined.

CMAKE_<lang>_LINK_PIE_SUPPORTED

Set to true if PIE is supported by the linker and false otherwise.

CMAKE_<lang>_LINK_NO_PIE_SUPPORTED

Set to true if NO_PIE is supported by the linker and false otherwise.

check_pie_supported()
set_property(TARGET foo PROPERTY POSITION_INDEPENDENT_CODE TRUE)

# Retrieve any error message.
check_pie_supported(OUTPUT_VARIABLE output LANGUAGES C)
set_property(TARGET foo PROPERTY POSITION_INDEPENDENT_CODE TRUE)
if(NOT CMAKE_C_LINK_PIE_SUPPORTED)
  message(WARNING "PIE is not supported at link time: ${output}.\n"
                  "PIE link options will not be passed to linker.")
endif()

Check if the prototype we expect is correct.

check_prototype_definition

check_prototype_definition(FUNCTION PROTOTYPE RETURN HEADER VARIABLE)

FUNCTION - The name of the function (used to check if prototype exists)
PROTOTYPE- The prototype to check.
RETURN - The return value of the function.
HEADER - The header files required.
VARIABLE - The variable to store the result.
           Will be created as an internal cache variable.

Example:

check_prototype_definition(getpwent_r
 "struct passwd *getpwent_r(struct passwd *src, char *buf, int buflen)"
 "NULL"
 "unistd.h;pwd.h"
 SOLARIS_GETPWENT_R)

The following variables may be set before calling this function to modify the way the check is run:

Added in version 3.19.

Check once if source code can be built for a given language.

check_source_compiles

check_source_compiles(<lang> <code> <resultVar>
                      [FAIL_REGEX <regex1> [<regex2>...]]
                      [SRC_EXT <extension>])

Check once that the source supplied in <code> can be built for code language <lang>. The result is stored in the internal cache variable specified by <resultVar>, with boolean true for success and boolean false for failure.

If FAIL_REGEX is provided, then failure is determined by checking if anything in the compiler output matches any of the specified regular expressions.

By default, the test source file will be given a file extension that matches the requested language. The SRC_EXT option can be used to override this with .<extension> instead.

The C example checks if the compiler supports the noreturn attribute:

set(CMAKE_TRY_COMPILE_TARGET_TYPE "STATIC_LIBRARY")

check_source_compiles(C
"#if !__has_c_attribute(noreturn)
#error \"No noreturn attribute\"
#endif"
HAVE_NORETURN)

The Fortran example checks if the compiler supports the pure procedure attribute:

set(CMAKE_TRY_COMPILE_TARGET_TYPE "STATIC_LIBRARY")

check_source_compiles(Fortran
"pure subroutine foo()
end subroutine"
HAVE_PURE)

Internally, try_compile() is used to compile the source. If CMAKE_TRY_COMPILE_TARGET_TYPE is set to EXECUTABLE (default), the source is compiled and linked as an executable program. If set to STATIC_LIBRARY, the source is compiled but not linked. In any case, all functions must be declared as usual.

See also check_source_runs() to run compiled source.

The compile and link commands can be influenced by setting any of the following variables prior to calling check_source_compiles():

Added in version 3.19.

Check if given source compiles and links into an executable and can subsequently be run.

check_source_runs

check_source_runs(<lang> <code> <resultVar>
                  [SRC_EXT <extension>])

Check once that the <lang> source supplied in <code> can be built, linked as an executable, and then run. The <code> must contain at least a main() function, or in Fortran a program.

The result is stored in the internal cache variable specified by <resultVar>. Success of build and run is indicated by boolean true. Failure to build or run is indicated by boolean false such as an empty string or an error message.

By default, the test source file will be given a file extension that matches the requested language. The SRC_EXT option can be used to override this with .<extension> instead.

The <code> must contain a valid main program. For example:

check_source_runs(C
"#include <stdlib.h>
#include <stdnoreturn.h>
noreturn void f(){ exit(0); }
int main(void) { f(); return 1; }"
HAVE_NORETURN)

check_source_runs(Fortran
"program test
real :: x[*]
call co_sum(x)
end program"
HAVE_COARRAY)

The compile and link commands can be influenced by setting any of the following variables prior to calling check_source_runs()

Check if the given struct or class has the specified member variable

CHECK_STRUCT_HAS_MEMBER

CHECK_STRUCT_HAS_MEMBER(<struct> <member> <header> <variable>
                        [LANGUAGE <language>])

<struct> - the name of the struct or class you are interested in
<member> - the member which existence you want to check
<header> - the header(s) where the prototype should be declared
<variable> - variable to store the result
<language> - the compiler to use (C or CXX)

The following variables may be set before calling this macro to modify the way the check is run:

Example:

CHECK_STRUCT_HAS_MEMBER("struct timeval" tv_sec sys/select.h
                        HAVE_TIMEVAL_TV_SEC LANGUAGE C)

Provides a macro to check if a symbol exists as a function, variable, or macro in C.

check_symbol_exists

check_symbol_exists(<symbol> <files> <variable>)

Check that the <symbol> is available after including given header <files> and store the result in a <variable>.  Specify the list of files in one argument as a semicolon-separated list. <variable> will be created as an internal cache variable.

If the header files define the symbol as a macro it is considered available and assumed to work.  If the header files declare the symbol as a function or variable then the symbol must also be available for linking (so intrinsics may not be detected). If the symbol is a type, enum value, or intrinsic it will not be recognized (consider using CheckTypeSize or CheckSourceCompiles). If the check needs to be done in C++, consider using CheckCXXSymbolExists instead.

The following variables may be set before calling this macro to modify the way the check is run:

For example:

include(CheckSymbolExists)

# Check for macro SEEK_SET
check_symbol_exists(SEEK_SET "stdio.h" HAVE_SEEK_SET)
# Check for function fopen
check_symbol_exists(fopen "stdio.h" HAVE_FOPEN)

Check sizeof a type

check_type_size

check_type_size(<type> <variable> [BUILTIN_TYPES_ONLY]
                                  [LANGUAGE <language>])

Check if the type exists and determine its size.  Results are reported in the following variables:

HAVE_<variable>

Holds a true or false value indicating whether the type exists.

<variable>

Holds one of the following values:

<size>

Type has non-zero size <size>.

0

Type has architecture-dependent size.  This may occur when CMAKE_OSX_ARCHITECTURES has multiple architectures. In this case <variable>_CODE contains C preprocessor tests mapping from each architecture macro to the corresponding type size. The list of architecture macros is stored in <variable>_KEYS, and the value for each key is stored in <variable>-<key>.

"" (empty string)

Type does not exist.

<variable>_CODE

Holds C preprocessor code to define the macro <variable> to the size of the type, or to leave the macro undefined if the type does not exist.

The options are:

BUILTIN_TYPES_ONLY

Support only compiler-builtin types.  If not given, the macro checks for headers <sys/types.h>, <stdint.h>, and <stddef.h>, and saves results in HAVE_SYS_TYPES_H, HAVE_STDINT_H, and HAVE_STDDEF_H.  The type size check automatically includes the available headers, thus supporting checks of types defined in the headers.

LANGUAGE <language>

Use the <language> compiler to perform the check. Acceptable values are C and CXX.

Despite the name of the macro you may use it to check the size of more complex expressions, too.  To check e.g.  for the size of a struct member you can do something like this:

check_type_size("((struct something*)0)->member" SIZEOF_MEMBER)

The following variables may be set before calling this macro to modify the way the check is run:

CMAKE_EXTRA_INCLUDE_FILES

list of extra headers to include.

Check if the variable exists.

CHECK_VARIABLE_EXISTS

CHECK_VARIABLE_EXISTS(VAR VARIABLE)

VAR      - the name of the variable
VARIABLE - variable to store the result
           Will be created as an internal cache variable.

This macro is only for C variables.

The following variables may be set before calling this macro to modify the way the check is run:

Add a fortran-only subdirectory, find a fortran compiler, and build.

The cmake_add_fortran_subdirectory function adds a subdirectory to a project that contains a fortran-only subproject.  The module will check the current compiler and see if it can support fortran.  If no fortran compiler is found and the compiler is MSVC, then this module will find the MinGW gfortran.  It will then use an external project to build with the MinGW tools.  It will also create imported targets for the libraries created.  This will only work if the fortran code is built into a dll, so BUILD_SHARED_LIBS is turned on in the project.  In addition the CMAKE_GNUtoMS option is set to on, so that Microsoft .lib files are created.  Usage is as follows:

cmake_add_fortran_subdirectory(
 <subdir>                # name of subdirectory
 PROJECT <project_name>  # project name in subdir top CMakeLists.txt
 ARCHIVE_DIR <dir>       # dir where project places .lib files
 RUNTIME_DIR <dir>       # dir where project places .dll files
 LIBRARIES <lib>...      # names of library targets to import
 LINK_LIBRARIES          # link interface libraries for LIBRARIES
  [LINK_LIBS <lib> <dep>...]...
 CMAKE_COMMAND_LINE ...  # extra command line flags to pass to cmake
 NO_EXTERNAL_INSTALL     # skip installation of external project
 )

Relative paths in ARCHIVE_DIR and RUNTIME_DIR are interpreted with respect to the build directory corresponding to the source directory in which the function is invoked.

Limitations:

NO_EXTERNAL_INSTALL is required for forward compatibility with a future version that supports installation of the external project binaries during make install.

define a bunch of backwards compatibility variables

CMAKE_ANSI_CXXFLAGS - flag for ansi c++
CMAKE_HAS_ANSI_STRING_STREAM - has <strstream>
include(TestForANSIStreamHeaders)
include(CheckIncludeFileCXX)
include(TestForSTDNamespace)
include(TestForANSIForScope)

Macro to provide an option dependent on other options.

This macro presents an option to the user only if a set of other conditions are true.

cmake_dependent_option

cmake_dependent_option(<option> "<help_text>" <value> <depends> <force>)

Makes <option> available to the user if the semicolon-separated list of conditions in <depends> are all true.  Otherwise, a local variable named <option> is set to <force>.

When <option> is available, the given <help_text> and initial <value> are used. Otherwise, any value set by the user is preserved for when <depends> is satisfied in the future.

Note that the <option> variable only has a value which satisfies the <depends> condition within the scope of the caller because it is a local variable.

Example invocation:

cmake_dependent_option(USE_FOO "Use Foo" ON "USE_BAR;NOT USE_ZOT" OFF)

If USE_BAR is true and USE_ZOT is false, this provides an option called USE_FOO that defaults to ON. Otherwise, it sets USE_FOO to OFF and hides the option from the user. If the status of USE_BAR or USE_ZOT ever changes, any value for the USE_FOO option is saved so that when the option is re-enabled it retains its old value.

Added in version 3.22: Full Condition Syntax is now supported.  See policy CMP0127.

find_dependency

The find_dependency() macro wraps a find_package() call for a package dependency:

find_dependency(<dep> [...])

It is designed to be used in a Package Configuration File (<PackageName>Config.cmake).  find_dependency forwards the correct parameters for QUIET and REQUIRED which were passed to the original find_package() call.  Any additional arguments specified are forwarded to find_package().

If the dependency could not be found it sets an informative diagnostic message and calls return() to end processing of the calling package configuration file and return to the find_package() command that loaded it.

NOTE:

The call to return() makes this macro unsuitable to call from Find Modules.

If find_dependency is called with arguments identical to a previous call in the same directory, perhaps due to diamond-shaped package dependencies, the underlying call to find_package() is optimized out.  This optimization is important to support large package dependency graphs while avoiding a combinatorial explosion of repeated searches. However, the heuristic cannot account for ambient variables that affect package behavior, such as <PackageName>_USE_STATIC_LIBS, offered by some packages.  Therefore package configuration files should avoid setting such variables before their calls to find_dependency.

Changed in version 3.15: Previously, the underlying call to find_package() was always optimized out if the package had already been found.  CMake 3.15 removed the optimization to support cases in which find_dependency call arguments request different components.

Changed in version 3.26: The pre-3.15 optimization was restored, but with the above-described heuristic to account for varying find_dependency call arguments.

This file is executed by cmake when invoked with --find-package. It expects that the following variables are set using -D:

NAME

name of the package

COMPILER_ID

the CMake compiler ID for which the result is, i.e. GNU/Intel/Clang/MSVC, etc.

LANGUAGE

language for which the result will be used, i.e. C/CXX/Fortran/ASM

MODE

EXIST

only check for existence of the given package

COMPILE

print the flags needed for compiling an object file which uses the given package

LINK

print the flags needed for linking when using the given package

QUIET

if TRUE, don't print anything

The builtin Graphviz support of CMake.

CMake can generate Graphviz files showing the dependencies between the targets in a project, as well as external libraries which are linked against.

When running CMake with the --graphviz=foo.dot option, it produces:

• a foo.dot file, showing all dependencies in the project
• a foo.dot.<target> file for each target, showing on which other targets it depends
• a foo.dot.<target>.dependers file for each target, showing which other targets depend on it

Those .dot files can be converted to images using the dot command from the Graphviz package:

dot -Tpng -o foo.png foo.dot

Added in version 3.10: The different dependency types PUBLIC, INTERFACE and PRIVATE are represented as solid, dashed and dotted edges.

The resulting graphs can be huge.  The look and content of the generated graphs can be controlled using the file CMakeGraphVizOptions.cmake.  This file is first searched in CMAKE_BINARY_DIR, and then in CMAKE_SOURCE_DIR.  If found, the variables set in it are used to adjust options for the generated Graphviz files.

GRAPHVIZ_GRAPH_NAME

The graph name.

• Mandatory: NO
• Default: value of CMAKE_PROJECT_NAME

GRAPHVIZ_GRAPH_HEADER

The header written at the top of the Graphviz files.

• Mandatory: NO
• Default: "node [ fontsize = "12" ];"

GRAPHVIZ_NODE_PREFIX

The prefix for each node in the Graphviz files.

• Mandatory: NO
• Default: "node"

GRAPHVIZ_EXECUTABLES

Set to FALSE to exclude executables from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_STATIC_LIBS

Set to FALSE to exclude static libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_SHARED_LIBS

Set to FALSE to exclude shared libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_MODULE_LIBS

Set to FALSE to exclude module libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_INTERFACE_LIBS

Set to FALSE to exclude interface libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_OBJECT_LIBS

Set to FALSE to exclude object libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_UNKNOWN_LIBS

Set to FALSE to exclude unknown libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_EXTERNAL_LIBS

Set to FALSE to exclude external libraries from the generated graphs.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_CUSTOM_TARGETS

Set to TRUE to include custom targets in the generated graphs.

• Mandatory: NO
• Default: FALSE

GRAPHVIZ_IGNORE_TARGETS

A list of regular expressions for names of targets to exclude from the generated graphs.

• Mandatory: NO
• Default: empty

GRAPHVIZ_GENERATE_PER_TARGET

Set to FALSE to not generate per-target graphs foo.dot.<target>.

• Mandatory: NO
• Default: TRUE

GRAPHVIZ_GENERATE_DEPENDERS

Set to FALSE to not generate depender graphs foo.dot.<target>.dependers.

• Mandatory: NO
• Default: TRUE

Helper functions for creating config files that can be included by other projects to find and use a package.

configure_package_config_file

Create a config file for a project:

configure_package_config_file(<input> <output>
  INSTALL_DESTINATION <path>
  [PATH_VARS <var1> <var2> ... <varN>]
  [NO_SET_AND_CHECK_MACRO]
  [NO_CHECK_REQUIRED_COMPONENTS_MACRO]
  [INSTALL_PREFIX <path>]
  )

configure_package_config_file() should be used instead of the plain configure_file() command when creating the <PackageName>Config.cmake or <PackageName>-config.cmake file for installing a project or library. It helps make the resulting package relocatable by avoiding hardcoded paths in the installed <PackageName>Config.cmake file.

In a FooConfig.cmake file there may be code like this to make the install destinations known to the using project:

set(FOO_INCLUDE_DIR   "@CMAKE_INSTALL_FULL_INCLUDEDIR@" )
set(FOO_DATA_DIR   "@CMAKE_INSTALL_PREFIX@/@RELATIVE_DATA_INSTALL_DIR@" )
set(FOO_ICONS_DIR   "@CMAKE_INSTALL_PREFIX@/share/icons" )
#...logic to determine installedPrefix from the own location...
set(FOO_CONFIG_DIR  "${installedPrefix}/@CONFIG_INSTALL_DIR@" )

All four options shown above are not sufficient  The first three hardcode the absolute directory locations.  The fourth case works only if the logic to determine the installedPrefix is correct, and if CONFIG_INSTALL_DIR contains a relative path, which in general cannot be guaranteed.  This has the effect that the resulting FooConfig.cmake file would work poorly under Windows and macOS, where users are used to choosing the install location of a binary package at install time, independent from how CMAKE_INSTALL_PREFIX was set at build/cmake time.

Using configure_package_config_file() helps.  If used correctly, it makes the resulting FooConfig.cmake file relocatable.  Usage:

1. Write a FooConfig.cmake.in file as you are used to.
2. Insert a line at the top containing only the string @PACKAGE_INIT@.
3. Instead of set(FOO_DIR "@SOME_INSTALL_DIR@"), use set(FOO_DIR "@PACKAGE_SOME_INSTALL_DIR@") (this must be after the @PACKAGE_INIT@ line).
4. Instead of using the normal configure_file() command, use configure_package_config_file().

The <input> and <output> arguments are the input and output file, the same way as in configure_file().

The <path> given to INSTALL_DESTINATION must be the destination where the FooConfig.cmake file will be installed to.  This path can either be absolute, or relative to the INSTALL_PREFIX path.

The variables <var1> to <varN> given as PATH_VARS are the variables which contain install destinations.  For each of them, the macro will create a helper variable PACKAGE_<var...>.  These helper variables must be used in the FooConfig.cmake.in file for setting the installed location. They are calculated by configure_package_config_file() so that they are always relative to the installed location of the package.  This works both for relative and also for absolute locations.  For absolute locations, it works only if the absolute location is a subdirectory of INSTALL_PREFIX.

Added in version 3.30: The variable PACKAGE_PREFIX_DIR will always be defined after the @PACKAGE_INIT@ line.  It will hold the value of the base install location.  In general, variables defined via the PATH_VARS mechanism should be used instead, but PACKAGE_PREFIX_DIR can be used for those cases not easily handled by PATH_VARS, such as for files installed directly to the base install location rather than a subdirectory of it.

NOTE:

Added in version 3.1: If the INSTALL_PREFIX argument is passed, this is used as the base path to calculate all the relative paths.  The <path> argument must be an absolute path.  If this argument is not passed, the CMAKE_INSTALL_PREFIX variable will be used instead.  The default value is good when generating a FooConfig.cmake file to use your package from the install tree.  When generating a FooConfig.cmake file to use your package from the build tree, this option should be used.

By default, configure_package_config_file() also generates two helper macros, set_and_check() and check_required_components(), into the FooConfig.cmake file.

set_and_check() should be used instead of the normal set() command for setting directories and file locations.  In addition to setting the variable, it also checks that the referenced file or directory actually exists and fails with a fatal error if it doesn't.  This ensures that the generated FooConfig.cmake file does not contain wrong references. Add the NO_SET_AND_CHECK_MACRO option to prevent the generation of the set_and_check() macro in the FooConfig.cmake file.

check_required_components(<PackageName>) should be called at the end of the FooConfig.cmake file. This macro checks whether all requested, non-optional components have been found, and if this is not the case, it sets the Foo_FOUND variable to FALSE so that the package is considered to be not found.  It does that by testing the Foo_<Component>_FOUND variables for all requested required components.  This macro should be called even if the package doesn't provide any components to make sure users are not specifying components erroneously.  Add the NO_CHECK_REQUIRED_COMPONENTS_MACRO option to prevent the generation of the check_required_components() macro in the FooConfig.cmake file.

See also Example Generating Package Files.

write_basic_package_version_file

Create a version file for a project:

write_basic_package_version_file(<filename>
  [VERSION <major.minor.patch>]
  COMPATIBILITY <AnyNewerVersion|SameMajorVersion|SameMinorVersion|ExactVersion>
  [ARCH_INDEPENDENT] )

Writes a file for use as a <PackageName>ConfigVersion.cmake file to <filename>.  See the documentation of find_package() for details on such files.

<filename> is the output filename, which should be in the build tree. <major.minor.patch> is the version number of the project to be installed.

If no VERSION is given, the PROJECT_VERSION variable is used. If this hasn't been set, it errors out.

The Compatibility mode AnyNewerVersion means that the installed package version will be considered compatible if it is newer or exactly the same as the requested version.  This mode should be used for packages which are fully backward compatible, also across major versions. If SameMajorVersion is used instead, then the behavior differs from AnyNewerVersion in that the major version number must be the same as requested, e.g.  version 2.0 will not be considered compatible if 1.0 is requested.  This mode should be used for packages which guarantee backward compatibility within the same major version. If SameMinorVersion is used, the behavior is the same as SameMajorVersion, but both major and minor version must be the same as requested, e.g version 0.2 will not be compatible if 0.1 is requested. If ExactVersion is used, then the package is only considered compatible if the requested version matches exactly its own version number (not considering the tweak version).  For example, version 1.2.3 of a package is only considered compatible to requested version 1.2.3.  This mode is for packages without compatibility guarantees. If your project has more elaborate version matching rules, you will need to write your own custom <PackageName>ConfigVersion.cmake file instead of using this macro.

Added in version 3.11: The SameMinorVersion compatibility mode.

Added in version 3.14: If ARCH_INDEPENDENT is given, the installed package version will be considered compatible even if it was built for a different architecture than the requested architecture.  Otherwise, an architecture check will be performed, and the package will be considered compatible only if the architecture matches exactly.  For example, if the package is built for a 32-bit architecture, the package is only considered compatible if it is used on a 32-bit architecture, unless ARCH_INDEPENDENT is given, in which case the package is considered compatible on any architecture.

NOTE:

Added in version 3.19: The version file generated by AnyNewerVersion, SameMajorVersion and SameMinorVersion arguments of Compatibility handle the version range, if one is specified (see find_package() command for the details). ExactVersion mode is incompatible with version ranges and will display an author warning if one is specified.

Internally, this macro executes configure_file() to create the resulting version file.  Depending on the Compatibility, the corresponding BasicConfigVersion-<Compatibility>.cmake.in file is used. Please note that these files are internal to CMake and you should not call configure_file() on them yourself, but they can be used as a starting point to create more sophisticated custom <PackageName>ConfigVersion.cmake files.

generate_apple_platform_selection_file

Added in version 3.29.

Create an Apple platform selection file:

generate_apple_platform_selection_file(<filename>
  INSTALL_DESTINATION <path>
  [INSTALL_PREFIX <path>]
  [MACOS_INCLUDE_FILE <file>]
  [IOS_INCLUDE_FILE <file>]
  [IOS_SIMULATOR_INCLUDE_FILE <file>]
  [IOS_CATALYST_INCLUDE_FILE <file>]
  [TVOS_INCLUDE_FILE <file>]
  [TVOS_SIMULATOR_INCLUDE_FILE <file>]
  [WATCHOS_INCLUDE_FILE <file>]
  [WATCHOS_SIMULATOR_INCLUDE_FILE <file>]
  [VISIONOS_INCLUDE_FILE <file>]
  [VISIONOS_SIMULATOR_INCLUDE_FILE <file>]
  [ERROR_VARIABLE <variable>]
  )

Write a file that includes an Apple-platform-specific .cmake file, e.g., for use as <PackageName>Config.cmake.  This can be used in conjunction with the XCFRAMEWORK_LOCATION argument of export(SETUP) to export packages in a way that a project built for any Apple platform can use them.

INSTALL_DESTINATION <path>

Path to which the generated file will be installed by the caller, e.g., via install(FILES).  The path may be either relative to the INSTALL_PREFIX or absolute.

INSTALL_PREFIX <path>

Path prefix to which the package will be installed by the caller. The <path> argument must be an absolute path.  If this argument is not passed, the CMAKE_INSTALL_PREFIX variable will be used instead.

MACOS_INCLUDE_FILE <file>

File to include if the platform is macOS.

IOS_INCLUDE_FILE <file>

File to include if the platform is iOS.

IOS_SIMULATOR_INCLUDE_FILE <file>

File to include if the platform is iOS Simulator.

IOS_CATALYST_INCLUDE_FILE <file>

Added in version 3.31.

File to include if the platform is iOS Catalyst.

TVOS_INCLUDE_FILE <file>

File to include if the platform is tvOS.

TVOS_SIMULATOR_INCLUDE_FILE <file>

File to include if the platform is tvOS Simulator.

WATCHOS_INCLUDE_FILE <file>

File to include if the platform is watchOS.

WATCHOS_SIMULATOR_INCLUDE_FILE <file>

File to include if the platform is watchOS Simulator.

VISIONOS_INCLUDE_FILE <file>

File to include if the platform is visionOS.

VISIONOS_SIMULATOR_INCLUDE_FILE <file>

File to include if the platform is visionOS Simulator.

ERROR_VARIABLE <variable>

If the consuming project is built for an unsupported platform, set <variable> to an error message.  The includer may use this information to pretend the package was not found.  If this option is not given, the default behavior is to issue a fatal error.

If any of the optional include files is not specified, and the consuming project is built for its corresponding platform, the generated file will consider the platform to be unsupported.  The behavior is determined by the ERROR_VARIABLE option.

generate_apple_architecture_selection_file

Added in version 3.29.

Create an Apple architecture selection file:

generate_apple_architecture_selection_file(<filename>
  INSTALL_DESTINATION <path>
  [INSTALL_PREFIX <path>]
  [SINGLE_ARCHITECTURES <arch>...
   SINGLE_ARCHITECTURE_INCLUDE_FILES <file>...]
  [UNIVERSAL_ARCHITECTURES <arch>...
   UNIVERSAL_INCLUDE_FILE <file>]
  [ERROR_VARIABLE <variable>]
  )

Write a file that includes an Apple-architecture-specific .cmake file based on CMAKE_OSX_ARCHITECTURES, e.g., for inclusion from an Apple-specific <PackageName>Config.cmake file.

INSTALL_DESTINATION <path>

Path to which the generated file will be installed by the caller, e.g., via install(FILES).  The path may be either relative to the INSTALL_PREFIX or absolute.

INSTALL_PREFIX <path>

Path prefix to which the package will be installed by the caller. The <path> argument must be an absolute path.  If this argument is not passed, the CMAKE_INSTALL_PREFIX variable will be used instead.

SINGLE_ARCHITECTURES <arch>...

Architectures provided by entries of SINGLE_ARCHITECTURE_INCLUDE_FILES.

SINGLE_ARCHITECTURE_INCLUDE_FILES <file>...

Architecture-specific files.  One of them will be loaded when CMAKE_OSX_ARCHITECTURES contains a single architecture matching the corresponding entry of SINGLE_ARCHITECTURES.

UNIVERSAL_ARCHITECTURES <arch>...

Architectures provided by the UNIVERSAL_INCLUDE_FILE.

The list may include $(ARCHS_STANDARD) to support consumption using the Xcode generator, but the architectures should always be listed individually too.

UNIVERSAL_INCLUDE_FILE <file>

A file to load when CMAKE_OSX_ARCHITECTURES contains a (non-strict) subset of the UNIVERSAL_ARCHITECTURES and does not match any one of the SINGLE_ARCHITECTURES.

ERROR_VARIABLE <variable>

If the consuming project is built for an unsupported architecture, set <variable> to an error message.  The includer may use this information to pretend the package was not found.  If this option is not given, the default behavior is to issue a fatal error.

Example using both the configure_package_config_file() and write_basic_package_version_file() commands:

CMakeLists.txt

include(GNUInstallDirs)
set(INCLUDE_INSTALL_DIR ${CMAKE_INSTALL_INCLUDEDIR}/Foo
    CACHE PATH "Location of header files" )
set(SYSCONFIG_INSTALL_DIR ${CMAKE_INSTALL_SYSCONFDIR}/foo
    CACHE PATH "Location of configuration files" )
#...
include(CMakePackageConfigHelpers)
configure_package_config_file(FooConfig.cmake.in
  ${CMAKE_CURRENT_BINARY_DIR}/FooConfig.cmake
  INSTALL_DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/Foo
  PATH_VARS INCLUDE_INSTALL_DIR SYSCONFIG_INSTALL_DIR)
write_basic_package_version_file(
  ${CMAKE_CURRENT_BINARY_DIR}/FooConfigVersion.cmake
  VERSION 1.2.3
  COMPATIBILITY SameMajorVersion )
install(FILES ${CMAKE_CURRENT_BINARY_DIR}/FooConfig.cmake
              ${CMAKE_CURRENT_BINARY_DIR}/FooConfigVersion.cmake
        DESTINATION ${CMAKE_INSTALL_LIBDIR}/cmake/Foo )

FooConfig.cmake.in

set(FOO_VERSION x.y.z)
...
@PACKAGE_INIT@
...
set_and_check(FOO_INCLUDE_DIR "@PACKAGE_INCLUDE_INSTALL_DIR@")
set_and_check(FOO_SYSCONFIG_DIR "@PACKAGE_SYSCONFIG_INSTALL_DIR@")

check_required_components(Foo)

Convenience functions for printing properties and variables, useful e.g. for debugging.

cmake_print_properties(<TARGETS       [<target1> ...] |
                        SOURCES       [<source1> ...] |
                        DIRECTORIES   [<dir1> ...]    |
                        TESTS         [<test1> ...]   |
                        CACHE_ENTRIES [<entry1> ...]  >
                       PROPERTIES [<prop1> ...]         )

This function prints the values of the properties of the given targets, source files, directories, tests or cache entries.  Exactly one of the scope keywords must be used.  The scope keyword and its arguments must come before the PROPERTIES keyword, in the arguments list.

Example:

cmake_print_properties(TARGETS foo bar PROPERTIES
                       LOCATION INTERFACE_INCLUDE_DIRECTORIES)

This will print the LOCATION and INTERFACE_INCLUDE_DIRECTORIES properties for both targets foo and bar.

cmake_print_variables(var1 var2 ..  varN)

This function will print the name of each variable followed by its value. Example:

cmake_print_variables(CMAKE_C_COMPILER CMAKE_MAJOR_VERSION DOES_NOT_EXIST)

Gives:

-- CMAKE_C_COMPILER="/usr/bin/gcc" ; CMAKE_MAJOR_VERSION="2" ; DOES_NOT_EXIST=""

Print system information.

This module serves diagnostic purposes. Just include it in a project to see various internal CMake variables.

This module defines three macros: CMAKE_PUSH_CHECK_STATE() CMAKE_POP_CHECK_STATE() and CMAKE_RESET_CHECK_STATE() These macros can be used to save, restore and reset (i.e., clear contents) the state of the variables CMAKE_REQUIRED_FLAGS, CMAKE_REQUIRED_DEFINITIONS, CMAKE_REQUIRED_LINK_OPTIONS, CMAKE_REQUIRED_LIBRARIES, CMAKE_REQUIRED_LINK_DIRECTORIES, CMAKE_REQUIRED_INCLUDES and CMAKE_EXTRA_INCLUDE_FILES used by the various Check-files coming with CMake, like e.g. check_function_exists() etc. The variable contents are pushed on a stack, pushing multiple times is supported.  This is useful e.g.  when executing such tests in a Find-module, where they have to be set, but after the Find-module has been executed they should have the same value as they had before.

CMAKE_PUSH_CHECK_STATE() macro receives optional argument RESET. Whether it's specified, CMAKE_PUSH_CHECK_STATE() will set all CMAKE_REQUIRED_* variables to empty values, same as CMAKE_RESET_CHECK_STATE() call will do.

Usage:

cmake_push_check_state(RESET)
set(CMAKE_REQUIRED_DEFINITIONS -DSOME_MORE_DEF)
check_function_exists(...)
cmake_reset_check_state()
set(CMAKE_REQUIRED_DEFINITIONS -DANOTHER_DEF)
check_function_exists(...)
cmake_pop_check_state()

CMakeVerifyManifest.cmake

This script is used to verify that embedded manifests and side by side manifests for a project match.  To run this script, cd to a directory and run the script with cmake -P.  On the command line you can pass in versions that are OK even if not found in the .manifest files.  For example, cmake -Dallow_versions=8.0.50608.0 -PCmakeVerifyManifest.cmake could be used to allow an embedded manifest of 8.0.50608.0 to be used in a project even if that version was not found in the .manifest file.

Configure generators for binary installers and source packages.

The CPack module generates the configuration files CPackConfig.cmake and CPackSourceConfig.cmake. They are intended for use in a subsequent run of  the cpack program where they steer the generation of installers or/and source packages.

Depending on the CMake generator, the CPack module may also add two new build targets, package and package_source. See the packaging targets section below for details.

The generated binary installers will contain all files that have been installed via CMake's install() command (and the deprecated commands install_files(), install_programs(), and install_targets()). Note that the DESTINATION option of the install() command must be a relative path; otherwise installed files are ignored by CPack.

Certain kinds of binary installers can be configured such that users can select individual application components to install.  See the CPackComponent module for further details.

Source packages (configured through CPackSourceConfig.cmake and generated by the CPack Archive Generator) will contain all source files in the project directory except those specified in CPACK_SOURCE_IGNORE_FILES.

The CPACK_GENERATOR variable has different meanings in different contexts.  In a CMakeLists.txt file, CPACK_GENERATOR is a list of generators: and when cpack is run with no other arguments, it will iterate over that list and produce one package for each generator.  In a CPACK_PROJECT_CONFIG_FILE, CPACK_GENERATOR is a string naming a single generator.  If you need per-cpack-generator logic to control other cpack settings, then you need a CPACK_PROJECT_CONFIG_FILE. If set, the CPACK_PROJECT_CONFIG_FILE is included automatically on a per-generator basis.  It only need contain overrides.

Here's how it works:

• cpack runs
• it includes CPackConfig.cmake
• it iterates over the generators given by the -G command line option, or if no such option was specified, over the list of generators given by the CPACK_GENERATOR variable set in the CPackConfig.cmake input file.

• foreach generator, it then

  • sets CPACK_GENERATOR to the one currently being iterated
  • includes the CPACK_PROJECT_CONFIG_FILE
  • produces the package for that generator

This is the key: For each generator listed in CPACK_GENERATOR in CPackConfig.cmake, cpack will reset CPACK_GENERATOR internally to the one currently being used and then include the CPACK_PROJECT_CONFIG_FILE.

For a list of available generators, see cpack-generators(7).

If CMake is run with the Makefile, Ninja, or Xcode generator, then include(CPack) generates a target package. This makes it possible to build a binary installer from CMake, Make, or Ninja: Instead of cpack, one may call cmake --build . --target package or make package or ninja package. The VS generator creates an uppercase target PACKAGE.

If CMake is run with the Makefile or Ninja generator, then include(CPack) also generates a target package_source. To build a source package, instead of cpack -G TGZ --config CPackSourceConfig.cmake one may call cmake --build . --target package_source, make package_source, or ninja package_source.

Before including this CPack module in your CMakeLists.txt file, there are a variety of variables that can be set to customize the resulting installers.  The most commonly-used variables are:

CPACK_PACKAGE_NAME

The name of the package (or application).  If not specified, it defaults to the project name.

CPACK_PACKAGE_VENDOR

The name of the package vendor. (e.g., "Kitware").  The default is "Humanity".

CPACK_PACKAGE_DIRECTORY

The directory in which CPack is doing its packaging.  If it is not set then this will default (internally) to the build dir.  This variable may be defined in a CPack config file or from the cpack command line option -B.  If set, the command line option overrides the value found in the config file.

CPACK_PACKAGE_VERSION_MAJOR

Package major version.  This variable will always be set, but its default value depends on whether or not version details were given to the project() command in the top level CMakeLists.txt file.  If version details were given, the default value will be CMAKE_PROJECT_VERSION_MAJOR.  If no version details were given, a default version of 0.1.1 will be assumed, leading to CPACK_PACKAGE_VERSION_MAJOR having a default value of 0.

CPACK_PACKAGE_VERSION_MINOR

Package minor version.  The default value is determined based on whether or not version details were given to the project() command in the top level CMakeLists.txt file.  If version details were given, the default value will be CMAKE_PROJECT_VERSION_MINOR, but if no minor version component was specified then CPACK_PACKAGE_VERSION_MINOR will be left unset.  If no project version was given at all, a default version of 0.1.1 will be assumed, leading to CPACK_PACKAGE_VERSION_MINOR having a default value of 1.

CPACK_PACKAGE_VERSION_PATCH

Package patch version.  The default value is determined based on whether or not version details were given to the project() command in the top level CMakeLists.txt file.  If version details were given, the default value will be CMAKE_PROJECT_VERSION_PATCH, but if no patch version component was specified then CPACK_PACKAGE_VERSION_PATCH will be left unset.  If no project version was given at all, a default version of 0.1.1 will be assumed, leading to CPACK_PACKAGE_VERSION_PATCH having a default value of 1.

CPACK_PACKAGE_DESCRIPTION

A description of the project, used in places such as the introduction screen of CPack-generated Windows installers.  If not set, the value of this variable is populated from the file named by CPACK_PACKAGE_DESCRIPTION_FILE.

CPACK_PACKAGE_DESCRIPTION_FILE

A text file used to describe the project when CPACK_PACKAGE_DESCRIPTION is not explicitly set.  The default value for CPACK_PACKAGE_DESCRIPTION_FILE points to a built-in template file Templates/CPack.GenericDescription.txt.

CPACK_PACKAGE_DESCRIPTION_SUMMARY

Short description of the project (only a few words).  If the CMAKE_PROJECT_DESCRIPTION variable is set, it is used as the default value, otherwise the default will be a string generated by CMake based on CMAKE_PROJECT_NAME.

CPACK_PACKAGE_HOMEPAGE_URL

Project homepage URL.  The default value is taken from the CMAKE_PROJECT_HOMEPAGE_URL variable, which is set by the top level project() command, or else the default will be empty if no URL was provided to project().

CPACK_PACKAGE_FILE_NAME

The name of the package file to generate, not including the extension.  For example, cmake-2.6.1-Linux-i686.  The default value is:

${CPACK_PACKAGE_NAME}-${CPACK_PACKAGE_VERSION}-${CPACK_SYSTEM_NAME}

CPACK_PACKAGE_INSTALL_DIRECTORY

Installation directory on the target system. This may be used by some CPack generators like NSIS to create an installation directory e.g., "CMake 2.5" below the installation prefix.  All installed elements will be put inside this directory.

CPACK_PACKAGE_ICON

A branding image that will be displayed inside the installer (used by GUI installers).

CPACK_PACKAGE_CHECKSUM

Added in version 3.7.

An algorithm that will be used to generate an additional file with the checksum of the package.  The output file name will be:

${CPACK_PACKAGE_FILE_NAME}.${CPACK_PACKAGE_CHECKSUM}

Supported algorithms are those listed by the string(<HASH>) command.

CPACK_PROJECT_CONFIG_FILE

CPack-time project CPack configuration file.  This file is included at cpack time, once per generator after CPack has set CPACK_GENERATOR to the actual generator being used.  It allows per-generator setting of CPACK_* variables at cpack time.

CPACK_RESOURCE_FILE_LICENSE

License to be embedded in the installer.  It will typically be displayed to the user by the produced installer (often with an explicit "Accept" button, for graphical installers) prior to installation.  This license file is NOT added to the installed files but is used by some CPack generators like NSIS.  If you want to use UTF-8 characters, the file needs to be encoded in UTF-8 BOM.  If you want to install a license file (may be the same as this one) along with your project, you must add an appropriate CMake install() command in your CMakeLists.txt.

CPACK_RESOURCE_FILE_README

ReadMe file to be embedded in the installer.  It typically describes in some detail the purpose of the project during the installation.  Not all CPack generators use this file.

CPACK_RESOURCE_FILE_WELCOME

Welcome file to be embedded in the installer.  It welcomes users to this installer.  Typically used in the graphical installers on Windows and Mac OS X.

CPACK_MONOLITHIC_INSTALL

Disables the component-based installation mechanism.  When set, the component specification is ignored and all installed items are put in a single "MONOLITHIC" package.  Some CPack generators do monolithic packaging by default and may be asked to do component packaging by setting CPACK_<GENNAME>_COMPONENT_INSTALL to TRUE.

CPACK_GENERATOR

List of CPack generators to use.  If not specified, CPack will create a set of options following the naming pattern CPACK_BINARY_<GENNAME> (e.g. CPACK_BINARY_NSIS) allowing the user to enable/disable individual generators.  If the -G option is given on the cpack command line, it will override this variable and any CPACK_BINARY_<GENNAME> options.

CPACK_OUTPUT_CONFIG_FILE

The name of the CPack binary configuration file.  This file is the CPack configuration generated by the CPack module for binary installers. Defaults to CPackConfig.cmake.

CPACK_PACKAGE_EXECUTABLES

Lists each of the executables and associated text label to be used to create Start Menu shortcuts.  For example, setting this to the list ccmake;CMake will create a shortcut named "CMake" that will execute the installed executable ccmake. Not all CPack generators use it (at least NSIS, Inno Setup and WIX do).

CPACK_STRIP_FILES

List of files to be stripped.  Starting with CMake 2.6.0, CPACK_STRIP_FILES will be a boolean variable which enables stripping of all files (a list of files evaluates to TRUE in CMake, so this change is compatible).

CPACK_VERBATIM_VARIABLES

Added in version 3.4.

If set to TRUE, values of variables prefixed with CPACK_ will be escaped before being written to the configuration files, so that the cpack program receives them exactly as they were specified.  If not, characters like quotes and backslashes can cause parsing errors or alter the value received by the cpack program.  Defaults to FALSE for backwards compatibility.

CPACK_THREADS

Added in version 3.20.

Number of threads to use when performing parallelized operations, such as compressing the installer package.

Some compression methods used by CPack generators such as Debian or Archive may take advantage of multiple CPU cores to speed up compression. CPACK_THREADS can be set to specify how many threads will be used for compression.

A positive integer can be used to specify an exact desired thread count.

When given a negative integer CPack will use the absolute value as the upper limit but may choose a lower value based on the available hardware concurrency.

Given 0 CPack will try to use all available CPU cores.

By default CPACK_THREADS is set to 1.

The following compression methods may take advantage of multiple cores:

xz

Supported if CMake is built with a liblzma that supports parallel compression.

Added in version 3.21: Official CMake binaries available on cmake.org now ship with a liblzma that supports parallel compression. Older versions did not.

zstd

Added in version 3.24.

Supported if CMake is built with libarchive 3.6 or higher. Official CMake binaries available on cmake.org support it.

Other compression methods ignore this value and use only one thread.

The following CPack variables are specific to source packages, and will not affect binary packages:

CPACK_SOURCE_PACKAGE_FILE_NAME

The name of the source package.  For example cmake-2.6.1.

CPACK_SOURCE_STRIP_FILES

List of files in the source tree that will be stripped.  Starting with CMake 2.6.0, CPACK_SOURCE_STRIP_FILES will be a boolean variable which enables stripping of all files (a list of files evaluates to TRUE in CMake, so this change is compatible).

CPACK_SOURCE_GENERATOR

List of generators used for the source packages.  As with CPACK_GENERATOR, if this is not specified then CPack will create a set of options (e.g. CPACK_SOURCE_ZIP) allowing users to select which packages will be generated.

CPACK_SOURCE_OUTPUT_CONFIG_FILE

The name of the CPack source configuration file.  This file is the CPack configuration generated by the CPack module for source installers. Defaults to CPackSourceConfig.cmake.

CPACK_SOURCE_IGNORE_FILES

Pattern of files in the source tree that won't be packaged when building a source package.  This is a list of regular expression patterns (that must be properly escaped), e.g., /CVS/;/\\.svn/;\\.swp$;\\.#;/#;.*~;cscope.*

The following variables are for advanced uses of CPack:

CPACK_CMAKE_GENERATOR

What CMake generator should be used if the project is a CMake project.  Defaults to the value of CMAKE_GENERATOR.  Few users will want to change this setting.

CPACK_INSTALL_CMAKE_PROJECTS

List of four values that specify what project to install.  The four values are: Build directory, Project Name, Project Component, Directory.  If omitted, CPack will build an installer that installs everything.

CPACK_SYSTEM_NAME

System name, defaults to the value of CMAKE_SYSTEM_NAME, except on Windows where it will be win32 or win64.

CPACK_PACKAGE_VERSION

Package full version, used internally.  By default, this is built from CPACK_PACKAGE_VERSION_MAJOR, CPACK_PACKAGE_VERSION_MINOR, and CPACK_PACKAGE_VERSION_PATCH.

CPACK_TOPLEVEL_TAG

Directory for the installed files.

CPACK_INSTALL_COMMANDS

Extra commands to install components.  The environment variable CMAKE_INSTALL_PREFIX is set to the temporary install directory during execution.

CPACK_INSTALL_SCRIPTS

Added in version 3.16.

Extra CMake scripts executed by CPack during its local staging installation.  They are executed before installing the files to be packaged. The scripts are not called by a standalone install (e.g.: make install). For every script, the following variables will be set: CMAKE_CURRENT_SOURCE_DIR, CMAKE_CURRENT_BINARY_DIR and CMAKE_INSTALL_PREFIX (which is set to the staging install directory).  The singular form CMAKE_INSTALL_SCRIPT is supported as an alternative variable for historical reasons, but its value is ignored if CMAKE_INSTALL_SCRIPTS is set and a warning will be issued.

See also CPACK_PRE_BUILD_SCRIPTS and CPACK_POST_BUILD_SCRIPTS which can be used to specify scripts to be executed later in the packaging process.

CPACK_PRE_BUILD_SCRIPTS

Added in version 3.19.

List of CMake scripts to execute after CPack has installed the files to be packaged into a staging directory and before producing the package(s) from those files. See also CPACK_INSTALL_SCRIPTS and CPACK_POST_BUILD_SCRIPTS.

CPACK_POST_BUILD_SCRIPTS

Added in version 3.19.

List of CMake scripts to execute after CPack has produced the resultant packages and before copying them back to the build directory. See also CPACK_INSTALL_SCRIPTS, CPACK_PRE_BUILD_SCRIPTS and CPACK_PACKAGE_FILES.

CPACK_PACKAGE_FILES

Added in version 3.19.

List of package files created in the staging directory, with each file provided as a full absolute path.  This variable is populated by CPack just before invoking the post-build scripts listed in CPACK_POST_BUILD_SCRIPTS.  It is the preferred way for the post-build scripts to know the set of package files to operate on. Projects should not try to set this variable themselves.

CPACK_INSTALLED_DIRECTORIES

Extra directories to install.

CPACK_PACKAGE_INSTALL_REGISTRY_KEY

Registry key used when installing this project.  This is only used by installers for Windows.  The default value is based on the installation directory.

CPACK_CREATE_DESKTOP_LINKS

List of desktop links to create.  Each desktop link requires a corresponding start menu shortcut as created by CPACK_PACKAGE_EXECUTABLES.

CPACK_BINARY_<GENNAME>

CPack generated options for binary generators.  The CPack.cmake module generates (when CPACK_GENERATOR is not set) a set of CMake options (see CMake option() command) which may then be used to select the CPack generator(s) to be used when building the package target or when running cpack without the -G option.

CPACK_READELF_EXECUTABLE

Added in version 3.25.

Specify the readelf executable path used by CPack. The default value will be taken from the CMAKE_READELF variable, if set, which may be populated by an internal CMake module.  If CMAKE_READELF is not set, CPack will use find_program() to determine the readelf path when needed.

CPACK_OBJCOPY_EXECUTABLE

Added in version 3.25.

Specify the objcopy executable path used by CPack. The default value will be taken from the CMAKE_OBJCOPY variable, if set, which may be populated by an internal CMake module.  If CMAKE_OBJCOPY is not set, CPack will use find_program() to determine the objcopy path when needed.

CPACK_OBJDUMP_EXECUTABLE

Added in version 3.25.

Specify the objdump executable path used by CPack. The default value will be taken from the CMAKE_OBJDUMP variable, if set, which may be populated by an internal CMake module.  If CMAKE_OBJDUMP is not set, CPack will use find_program() to determine the objdump path when needed.

Configure components for binary installers and source packages.

This module is automatically included by CPack.

Certain binary installers (especially the graphical installers) generated by CPack allow users to select individual application components to install. This module allows developers to configure the packaging of such components.

Contents is assigned to components by the COMPONENT argument of CMake's install() command.  Components can be annotated with user-friendly names and descriptions, inter-component dependencies, etc., and grouped in various ways to customize the resulting installer, using the commands described below.

To specify different groupings for different CPack generators use a CPACK_PROJECT_CONFIG_FILE.

The following variables influence the component-specific packaging:

CPACK_COMPONENTS_ALL

The list of component to install.

The default value of this variable is computed by CPack and contains all components defined by the project.  The user may set it to only include the specified components.

Instead of specifying all the desired components, it is possible to obtain a list of all defined components and then remove the unwanted ones from the list. The get_cmake_property() command can be used to obtain the Components property, then the list(REMOVE_ITEM) command can be used to remove the unwanted ones. For example, to use all defined components except foo and bar:

get_cmake_property(CPACK_COMPONENTS_ALL COMPONENTS)
list(REMOVE_ITEM CPACK_COMPONENTS_ALL "foo" "bar")

CPACK_<GENNAME>_COMPONENT_INSTALL

Enable/Disable component install for CPack generator <GENNAME>.

Each CPack Generator (RPM, DEB, ARCHIVE, NSIS, DMG, etc...) has a legacy default behavior.  e.g.  RPM builds monolithic whereas NSIS builds component.  One can change the default behavior by setting this variable to 0/1 or OFF/ON.

CPACK_COMPONENTS_GROUPING

Specify how components are grouped for multi-package component-aware CPack generators.

Some generators like RPM or ARCHIVE (TGZ, ZIP, ...) may generate several packages files when there are components, depending on the value of this variable:

• ONE_PER_GROUP (default): create one package per component group
• IGNORE : create one package per component (ignore the groups)
• ALL_COMPONENTS_IN_ONE : create a single package with all requested components

CPACK_COMPONENT_<compName>_DISPLAY_NAME

The name to be displayed for a component.

CPACK_COMPONENT_<compName>_DESCRIPTION

The description of a component.

CPACK_COMPONENT_<compName>_GROUP

The group of a component.

CPACK_COMPONENT_<compName>_DEPENDS

The dependencies (list of components) on which this component depends.

CPACK_COMPONENT_<compName>_HIDDEN

True if this component is hidden from the user.

CPACK_COMPONENT_<compName>_REQUIRED

True if this component is required.

CPACK_COMPONENT_<compName>_DISABLED

True if this component is not selected to be installed by default.

cpack_add_component

Describe an installation component.

cpack_add_component(compname
                    [DISPLAY_NAME name]
                    [DESCRIPTION description]
                    [HIDDEN | REQUIRED | DISABLED ]
                    [GROUP group]
                    [DEPENDS comp1 comp2 ... ]
                    [INSTALL_TYPES type1 type2 ... ]
                    [DOWNLOADED]
                    [ARCHIVE_FILE filename]
                    [PLIST filename])

compname is the name of an installation component, as defined by the COMPONENT argument of one or more CMake install() commands. With the cpack_add_component command one can set a name, a description, and other attributes of an installation component. One can also assign a component to a component group.

DISPLAY_NAME is the displayed name of the component, used in graphical installers to display the component name.  This value can be any string.

DESCRIPTION is an extended description of the component, used in graphical installers to give the user additional information about the component.  Descriptions can span multiple lines using \n as the line separator.  Typically, these descriptions should be no more than a few lines long.

HIDDEN indicates that this component will be hidden in the graphical installer, so that the user cannot directly change whether it is installed or not.

REQUIRED indicates that this component is required, and therefore will always be installed.  It will be visible in the graphical installer, but it cannot be unselected.  (Typically, required components are shown grayed out).

DISABLED indicates that this component should be disabled (unselected) by default.  The user is free to select this component for installation, unless it is also HIDDEN.

DEPENDS lists the components on which this component depends.  If this component is selected, then each of the components listed must also be selected.  The dependency information is encoded within the installer itself, so that users cannot install inconsistent sets of components.

GROUP names the component group of which this component is a part.  If not provided, the component will be a standalone component, not part of any component group.  Component groups are described with the cpack_add_component_group command, detailed below.

INSTALL_TYPES lists the installation types of which this component is a part.  When one of these installations types is selected, this component will automatically be selected.  Installation types are described with the cpack_add_install_type command, detailed below.

DOWNLOADED indicates that this component should be downloaded on-the-fly by the installer, rather than packaged in with the installer itself.  For more information, see the cpack_configure_downloads command.

ARCHIVE_FILE provides a name for the archive file created by CPack to be used for downloaded components.  If not supplied, CPack will create a file with some name based on CPACK_PACKAGE_FILE_NAME and the name of the component.  See cpack_configure_downloads for more information.

PLIST gives a filename that is passed to pkgbuild with the --component-plist argument when using the productbuild generator.

cpack_add_component_group

Describes a group of related CPack installation components.

cpack_add_component_group(groupname
                         [DISPLAY_NAME name]
                         [DESCRIPTION description]
                         [PARENT_GROUP parent]
                         [EXPANDED]
                         [BOLD_TITLE])

The cpack_add_component_group describes a group of installation components, which will be placed together within the listing of options.  Typically, component groups allow the user to select/deselect all of the components within a single group via a single group-level option.  Use component groups to reduce the complexity of installers with many options.  groupname is an arbitrary name used to identify the group in the GROUP argument of the cpack_add_component command, which is used to place a component in a group.  The name of the group must not conflict with the name of any component.

DISPLAY_NAME is the displayed name of the component group, used in graphical installers to display the component group name.  This value can be any string.

DESCRIPTION is an extended description of the component group, used in graphical installers to give the user additional information about the components within that group.  Descriptions can span multiple lines using \n as the line separator.  Typically, these descriptions should be no more than a few lines long.

PARENT_GROUP, if supplied, names the parent group of this group. Parent groups are used to establish a hierarchy of groups, providing an arbitrary hierarchy of groups.

EXPANDED indicates that, by default, the group should show up as "expanded", so that the user immediately sees all of the components within the group.  Otherwise, the group will initially show up as a single entry.

BOLD_TITLE indicates that the group title should appear in bold, to call the user's attention to the group.

cpack_add_install_type

Add a new installation type containing a set of predefined component selections to the graphical installer.

cpack_add_install_type(typename
                       [DISPLAY_NAME name])

The cpack_add_install_type command identifies a set of preselected components that represents a common use case for an application.  For example, a "Developer" install type might include an application along with its header and library files, while an "End user" install type might just include the application's executable.  Each component identifies itself with one or more install types via the INSTALL_TYPES argument to cpack_add_component.

DISPLAY_NAME is the displayed name of the install type, which will typically show up in a drop-down box within a graphical installer. This value can be any string.

cpack_configure_downloads

Configure CPack to download selected components on-the-fly as part of the installation process.

cpack_configure_downloads(site
                          [UPLOAD_DIRECTORY dirname]
                          [ALL]
                          [ADD_REMOVE|NO_ADD_REMOVE])

The cpack_configure_downloads command configures installation-time downloads of selected components.  For each downloadable component, CPack will create an archive containing the contents of that component, which should be uploaded to the given site.  When the user selects that component for installation, the installer will download and extract the component in place.  This feature is useful for creating small installers that only download the requested components, saving bandwidth.  Additionally, the installers are small enough that they will be installed as part of the normal installation process, and the "Change" button in Windows Add/Remove Programs control panel will allow one to add or remove parts of the application after the original installation.  On Windows, the downloaded-components functionality requires the ZipDLL plug-in for NSIS, available at:

http://nsis.sourceforge.net/ZipDLL_plug-in

On macOS, installers that download components on-the-fly can only be built and installed on system using macOS 10.5 or later.

The site argument is a URL where the archives for downloadable components will reside, e.g., https://cmake.org/files/v3.25/ All of the archives produced by CPack should be uploaded to that location.

UPLOAD_DIRECTORY is the local directory where CPack will create the various archives for each of the components.  The contents of this directory should be uploaded to a location accessible by the URL given in the site argument.  If omitted, CPack will use the directory CPackUploads inside the CMake binary directory to store the generated archives.

The ALL flag indicates that all components be downloaded.  Otherwise, only those components explicitly marked as DOWNLOADED or that have a specified ARCHIVE_FILE will be downloaded.  Additionally, the ALL option implies ADD_REMOVE (unless NO_ADD_REMOVE is specified).

ADD_REMOVE indicates that CPack should install a copy of the installer that can be called from Windows' Add/Remove Programs dialog (via the "Modify" button) to change the set of installed components. NO_ADD_REMOVE turns off this behavior.  This option is ignored on Mac OS X.

Added in version 3.1.

This module looks for the location of the command-line utilities supplied with the Qt Installer Framework (QtIFW).

The module also defines several commands to control the behavior of the CPack IFW Generator.

The module defines the following commands:

cpack_ifw_configure_component

Sets the arguments specific to the CPack IFW generator.

cpack_ifw_configure_component(<compname> [COMMON] [ESSENTIAL] [VIRTUAL]
                    [FORCED_INSTALLATION] [REQUIRES_ADMIN_RIGHTS]
                    [NAME <name>]
                    [DISPLAY_NAME <display_name>] # Note: Internationalization supported
                    [DESCRIPTION <description>] # Note: Internationalization supported
                    [UPDATE_TEXT <update_text>]
                    [VERSION <version>]
                    [RELEASE_DATE <release_date>]
                    [SCRIPT <script>]
                    [PRIORITY|SORTING_PRIORITY <sorting_priority>] # Note: PRIORITY is deprecated
                    [DEPENDS|DEPENDENCIES <com_id> ...]
                    [AUTO_DEPEND_ON <comp_id> ...]
                    [LICENSES <display_name> <file_path> ...]
                    [DEFAULT <value>]
                    [USER_INTERFACES <file_path> <file_path> ...]
                    [TRANSLATIONS <file_path> <file_path> ...]
                    [REPLACES <comp_id> ...]
                    [CHECKABLE <value>])

This command should be called after cpack_add_component() command.

COMMON

if set, then the component will be packaged and installed as part of a group to which it belongs.

ESSENTIAL

Added in version 3.6.

if set, then the package manager stays disabled until that component is updated.

VIRTUAL

Added in version 3.8.

if set, then the component will be hidden from the installer. It is a equivalent of the HIDDEN option from the cpack_add_component() command.

FORCED_INSTALLATION

Added in version 3.8.

if set, then the component must always be installed. It is a equivalent of the REQUIRED option from the cpack_add_component() command.

REQUIRES_ADMIN_RIGHTS

Added in version 3.8.

set it if the component needs to be installed with elevated permissions.

NAME

is used to create domain-like identification for this component. By default used origin component name.

DISPLAY_NAME

Added in version 3.8.

set to rewrite original name configured by cpack_add_component() command.

DESCRIPTION

Added in version 3.8.

set to rewrite original description configured by cpack_add_component() command.

UPDATE_TEXT

Added in version 3.8.

will be added to the component description if this is an update to the component.

VERSION

is version of component. By default used CPACK_PACKAGE_VERSION.

RELEASE_DATE

Added in version 3.8.

keep empty to auto generate.

SCRIPT

is a relative or absolute path to operations script for this component.

SORTING_PRIORITY

Added in version 3.8.

is priority of the component in the tree.

PRIORITY

Deprecated since version 3.8: Old name for SORTING_PRIORITY.

DEPENDS, DEPENDENCIES

Added in version 3.8.

list of dependency component or component group identifiers in QtIFW style.

Added in version 3.21.

Component or group names listed as dependencies may contain hyphens. This requires QtIFW 3.1 or later.

AUTO_DEPEND_ON

Added in version 3.8.

list of identifiers of component or component group in QtIFW style that this component has an automatic dependency on.

LICENSES

pair of <display_name> and <file_path> of license text for this component. You can specify more then one license.

DEFAULT

Added in version 3.8.

Possible values are: TRUE, FALSE, and SCRIPT. Set to FALSE to disable the component in the installer or to SCRIPT to resolved during runtime (don't forget add the file of the script as a value of the SCRIPT option).

USER_INTERFACES

Added in version 3.7.

is a list of <file_path> ('.ui' files) representing pages to load.

TRANSLATIONS

Added in version 3.8.

is a list of <file_path> ('.qm' files) representing translations to load.

REPLACES

Added in version 3.10.

list of identifiers of component or component group to replace.

CHECKABLE

Added in version 3.10.

Possible values are: TRUE, FALSE. Set to FALSE if you want to hide the checkbox for an item. This is useful when only a few subcomponents should be selected instead of all.

cpack_ifw_configure_component_group

Sets the arguments specific to the CPack IFW generator.

cpack_ifw_configure_component_group(<groupname> [VIRTUAL]
                    [FORCED_INSTALLATION] [REQUIRES_ADMIN_RIGHTS]
                    [NAME <name>]
                    [DISPLAY_NAME <display_name>] # Note: Internationalization supported
                    [DESCRIPTION <description>] # Note: Internationalization supported
                    [UPDATE_TEXT <update_text>]
                    [VERSION <version>]
                    [RELEASE_DATE <release_date>]
                    [SCRIPT <script>]
                    [PRIORITY|SORTING_PRIORITY <sorting_priority>] # Note: PRIORITY is deprecated
                    [DEPENDS|DEPENDENCIES <com_id> ...]
                    [AUTO_DEPEND_ON <comp_id> ...]
                    [LICENSES <display_name> <file_path> ...]
                    [DEFAULT <value>]
                    [USER_INTERFACES <file_path> <file_path> ...]
                    [TRANSLATIONS <file_path> <file_path> ...]
                    [REPLACES <comp_id> ...]
                    [CHECKABLE <value>])

This command should be called after cpack_add_component_group() command.

VIRTUAL

Added in version 3.8.

if set, then the group will be hidden from the installer. Note that setting this on a root component does not work.

FORCED_INSTALLATION

Added in version 3.8.

if set, then the group must always be installed.

REQUIRES_ADMIN_RIGHTS

Added in version 3.8.

set it if the component group needs to be installed with elevated permissions.

NAME

is used to create domain-like identification for this component group. By default used origin component group name.

DISPLAY_NAME

Added in version 3.8.

set to rewrite original name configured by cpack_add_component_group() command.

DESCRIPTION

Added in version 3.8.

set to rewrite original description configured by cpack_add_component_group() command.

UPDATE_TEXT

Added in version 3.8.

will be added to the component group description if this is an update to the component group.

VERSION

is version of component group. By default used CPACK_PACKAGE_VERSION.

RELEASE_DATE

Added in version 3.8.

keep empty to auto generate.

SCRIPT

is a relative or absolute path to operations script for this component group.

SORTING_PRIORITY

is priority of the component group in the tree.

PRIORITY

Deprecated since version 3.8: Old name for SORTING_PRIORITY.

DEPENDS, DEPENDENCIES

Added in version 3.8.

list of dependency component or component group identifiers in QtIFW style.

Added in version 3.21.

Component or group names listed as dependencies may contain hyphens. This requires QtIFW 3.1 or later.

AUTO_DEPEND_ON

Added in version 3.8.

list of identifiers of component or component group in QtIFW style that this component group has an automatic dependency on.

LICENSES

pair of <display_name> and <file_path> of license text for this component group. You can specify more then one license.

DEFAULT

Added in version 3.8.

Possible values are: TRUE, FALSE, and SCRIPT. Set to TRUE to preselect the group in the installer (this takes effect only on groups that have no visible child components) or to SCRIPT to resolved during runtime (don't forget add the file of the script as a value of the SCRIPT option).

USER_INTERFACES

Added in version 3.7.

is a list of <file_path> ('.ui' files) representing pages to load.

TRANSLATIONS

Added in version 3.8.

is a list of <file_path> ('.qm' files) representing translations to load.

REPLACES

Added in version 3.10.

list of identifiers of component or component group to replace.

CHECKABLE

Added in version 3.10.

Possible values are: TRUE, FALSE. Set to FALSE if you want to hide the checkbox for an item. This is useful when only a few subcomponents should be selected instead of all.

cpack_ifw_add_repository

Add QtIFW specific remote repository to binary installer.

cpack_ifw_add_repository(<reponame> [DISABLED]
                    URL <url>
                    [USERNAME <username>]
                    [PASSWORD <password>]
                    [DISPLAY_NAME <display_name>])

This command will also add the <reponame> repository to a variable CPACK_IFW_REPOSITORIES_ALL.

DISABLED

if set, then the repository will be disabled by default.

URL

is points to a list of available components.

USERNAME

is used as user on a protected repository.

PASSWORD

is password to use on a protected repository.

DISPLAY_NAME

is string to display instead of the URL.

cpack_ifw_update_repository

Added in version 3.6.

Update QtIFW specific repository from remote repository.

cpack_ifw_update_repository(<reponame>
                    [[ADD|REMOVE] URL <url>]|
                     [REPLACE OLD_URL <old_url> NEW_URL <new_url>]]
                    [USERNAME <username>]
                    [PASSWORD <password>]
                    [DISPLAY_NAME <display_name>])

This command will also add the <reponame> repository to a variable CPACK_IFW_REPOSITORIES_ALL.

URL

is points to a list of available components.

OLD_URL

is points to a list that will replaced.

NEW_URL

is points to a list that will replace to.

USERNAME

is used as user on a protected repository.

PASSWORD

is password to use on a protected repository.

DISPLAY_NAME

is string to display instead of the URL.

cpack_ifw_add_package_resources

Added in version 3.7.

Add additional resources in the installer binary.

cpack_ifw_add_package_resources(<file_path> <file_path> ...)

This command will also add the specified files to a variable CPACK_IFW_PACKAGE_RESOURCES.

Added in version 3.8.

The module defines configure_file() similar command to configure file templates prepared in QtIFW/SDK/Creator style.

The module defines the following commands:

cpack_ifw_configure_file

Copy a file to another location and modify its contents.

cpack_ifw_configure_file(<input> <output>)

Copies an <input> file to an <output> file and substitutes variable values referenced as %{VAR} or %VAR% in the input file content. Each variable reference will be replaced with the current value of the variable, or the empty string if the variable is not defined.

Added in version 3.8.

Functions to make configuration of CSharp/.NET targets easier.

A collection of CMake utility functions useful for dealing with CSharp targets for Visual Studio generators from version 2010 and later.

The following functions are provided by this module:

Main functions

• csharp_set_windows_forms_properties()
• csharp_set_designer_cs_properties()
• csharp_set_xaml_cs_properties()

Helper functions

• csharp_get_filename_keys()
• csharp_get_filename_key_base()
• csharp_get_dependentupon_name()

csharp_set_windows_forms_properties

Sets source file properties for use of Windows Forms. Use this, if your CSharp target uses Windows Forms:

csharp_set_windows_forms_properties([<file1> [<file2> [...]]])

<fileN>

List of all source files which are relevant for setting the VS_CSHARP_<tagname> properties (including .cs, .resx and .Designer.cs extensions).

In the list of all given files for all files ending with .Designer.cs and .resx is searched.  For every designer or resource file a file with the same base name but only .cs as extension is searched.  If this is found, the VS_CSHARP_<tagname> properties are set as follows:

for the .cs file:

• VS_CSHARP_SubType "Form"

for the .Designer.cs file (if it exists):

• VS_CSHARP_DependentUpon <cs-filename>
• VS_CSHARP_DesignTime "" (delete tag if previously defined)
• VS_CSHARP_AutoGen ""(delete tag if previously defined)

for the .resx file (if it exists):

• VS_RESOURCE_GENERATOR "" (delete tag if previously defined)
• VS_CSHARP_DependentUpon <cs-filename>
• VS_CSHARP_SubType "Designer"

csharp_set_designer_cs_properties

Sets source file properties of .Designer.cs files depending on sibling filenames. Use this, if your CSharp target does not use Windows Forms (for Windows Forms use csharp_set_windows_forms_properties() instead):

csharp_set_designer_cs_properties([<file1> [<file2> [...]]])

<fileN>

List of all source files which are relevant for setting the VS_CSHARP_<tagname> properties (including .cs, .resx, .settings and .Designer.cs extensions).

In the list of all given files for all files ending with .Designer.cs is searched. For every designer file all files with the same base name but different extensions are searched. If a match is found, the source file properties of the designer file are set depending on the extension of the matched file:

if match is .resx file:

• VS_CSHARP_AutoGen "True"
• VS_CSHARP_DesignTime "True"
• VS_CSHARP_DependentUpon <resx-filename>

if match is .cs file:

• VS_CSHARP_DependentUpon <cs-filename>

if match is .settings file:

• VS_CSHARP_AutoGen "True"
• VS_CSHARP_DesignTimeSharedInput "True"
• VS_CSHARP_DependentUpon <settings-filename>

NOTE:

Because the source file properties of the .Designer.cs file are set according to the found matches and every match sets the VS_CSHARP_DependentUpon property, there should only be one match for each Designer.cs file.

csharp_set_xaml_cs_properties

Sets source file properties for use of Windows Presentation Foundation (WPF) and XAML. Use this, if your CSharp target uses WPF/XAML:

csharp_set_xaml_cs_properties([<file1> [<file2> [...]]])

<fileN>

List of all source files which are relevant for setting the VS_CSHARP_<tagname> properties (including .cs, .xaml, and .xaml.cs extensions).

In the list of all given files for all files ending with .xaml.cs is searched. For every xaml-cs file, a file with the same base name but extension .xaml is searched. If a match is found, the source file properties of the .xaml.cs file are set:

• VS_CSHARP_DependentUpon <xaml-filename>

csharp_get_filename_keys

Helper function which computes a list of key values to identify source files independently of relative/absolute paths given in cmake and eliminates case sensitivity:

csharp_get_filename_keys(OUT [<file1> [<file2> [...]]])

OUT

Name of the variable in which the list of keys is stored

<fileN>

filename(s) as given to CSharp target using add_library() or add_executable()

In some way the function applies a canonicalization to the source names. This is necessary to find file matches if the files have been added to the target with different directory prefixes:

add_library(lib
  myfile.cs
  ${CMAKE_CURRENT_SOURCE_DIR}/myfile.Designer.cs)

set_source_files_properties(myfile.Designer.cs PROPERTIES
  VS_CSHARP_DependentUpon myfile.cs)

# this will fail, because in cmake
#  - ${CMAKE_CURRENT_SOURCE_DIR}/myfile.Designer.cs
#  - myfile.Designer.cs
# are not the same source file. The source file property is not set.

csharp_get_filename_key_base

Returns the full filepath and name without extension of a key. KEY is expected to be a key from csharp_get_filename_keys. In BASE the value of KEY without the file extension is returned:

csharp_get_filename_key_base(BASE KEY)

BASE

Name of the variable with the computed "base" of KEY.

KEY

The key of which the base will be computed. Expected to be a upper case full filename.

csharp_get_dependentupon_name

Computes a string which can be used as value for the source file property VS_CSHARP_<tagname> with target being DependentUpon:

csharp_get_dependentupon_name(NAME FILE)

NAME

Name of the variable with the result value

FILE

Filename to convert to <DependentUpon> value

Actually this is only the filename without any path given at the moment.

Configure a project for testing with CTest/CDash

Include this module in the top CMakeLists.txt file of a project to enable testing with CTest and dashboard submissions to CDash:

project(MyProject)
...
include(CTest)

The module automatically creates a BUILD_TESTING option that selects whether to enable testing support (ON by default).  After including the module, use code like:

if(BUILD_TESTING)
  # ... CMake code to create tests ...
endif()

to creating tests when testing is enabled.

To enable submissions to a CDash server, create a CTestConfig.cmake file at the top of the project with content such as:

set(CTEST_NIGHTLY_START_TIME "01:00:00 UTC")
set(CTEST_SUBMIT_URL "http://my.cdash.org/submit.php?project=MyProject")

(the CDash server can provide the file to a project administrator who configures MyProject).  Settings in the config file are shared by both this CTest module and the ctest(1) command-line Dashboard Client mode (ctest -S).

While building a project for submission to CDash, CTest scans the build output for errors and warnings and reports them with surrounding context from the build log.  This generic approach works for all build tools, but does not give details about the command invocation that produced a given problem.  One may get more detailed reports by setting the CTEST_USE_LAUNCHERS variable:

set(CTEST_USE_LAUNCHERS 1)

in the CTestConfig.cmake file.

Added in version 3.2.

This module provides the ctest_coverage_collect_gcov function.

This function runs gcov on all .gcda files found in the binary tree and packages the resulting .gcov files into a tar file. This tarball also contains the following:

• data.json defines the source and build directories for use by CDash.
• Labels.json indicates any LABELS that have been set on the source files.
• The uncovered directory holds any uncovered files found by CTEST_EXTRA_COVERAGE_GLOB.

After generating this tar file, it can be sent to CDash for display with the ctest_submit(CDASH_UPLOAD) command.

ctest_coverage_collect_gcov

ctest_coverage_collect_gcov(TARBALL <tarfile>
  [SOURCE <source_dir>][BUILD <build_dir>]
  [GCOV_COMMAND <gcov_command>]
  [GCOV_OPTIONS <options>...]
  )

Run gcov and package a tar file for CDash.  The options are:

TARBALL <tarfile>

Specify the location of the .tar file to be created for later upload to CDash.  Relative paths will be interpreted with respect to the top-level build directory.

TARBALL_COMPRESSION <option>

Added in version 3.18.

Specify a compression algorithm for the TARBALL data file.  Using this option reduces the size of the data file before it is submitted to CDash.  <option> must be one of GZIP, BZIP2, XZ, ZSTD, FROM_EXT, or an expression that CMake evaluates as FALSE. The default value is BZIP2.

If FROM_EXT is specified, the resulting file will be compressed based on the file extension of the <tarfile> (i.e. .tar.gz will use GZIP compression). File extensions that will produce compressed output include .tar.gz, .tgz, .tar.bzip2, .tbz, .tar.xz, and .txz.

SOURCE <source_dir>

Specify the top-level source directory for the build. Default is the value of CTEST_SOURCE_DIRECTORY.

BUILD <build_dir>

Specify the top-level build directory for the build. Default is the value of CTEST_BINARY_DIRECTORY.

GCOV_COMMAND <gcov_command>

Specify the full path to the gcov command on the machine. Default is the value of CTEST_COVERAGE_COMMAND.

GCOV_OPTIONS <options>...

Specify options to be passed to gcov.  The gcov command is run as gcov <options>... -o <gcov-dir> <file>.gcda. If not specified, the default option is just -b -x.

GLOB

Added in version 3.6.

Recursively search for .gcda files in build_dir rather than determining search locations by reading TargetDirectories.txt.

DELETE

Added in version 3.6.

Delete coverage files after they've been packaged into the .tar.

QUIET

Suppress non-error messages that otherwise would have been printed out by this function.

Added in version 3.3: Added support for the CTEST_CUSTOM_COVERAGE_EXCLUDE variable.

This file is read by ctest in script mode (-S)

Set the RULE_LAUNCH_* global properties when CTEST_USE_LAUNCHERS is on.

CTestUseLaunchers is automatically included when you include(CTest). However, it is split out into its own module file so projects can use the CTEST_USE_LAUNCHERS functionality independently.

To use launchers, set CTEST_USE_LAUNCHERS to ON in a ctest -S dashboard script, and then also set it in the cache of the configured project.  Both cmake and ctest need to know the value of it for the launchers to work properly.  CMake needs to know in order to generate proper build rules, and ctest, in order to produce the proper error and warning analysis.

For convenience, you may set the ENV variable CTEST_USE_LAUNCHERS_DEFAULT in your ctest -S script, too.  Then, as long as your CMakeLists uses include(CTest) or include(CTestUseLaunchers), it will use the value of the ENV variable to initialize a CTEST_USE_LAUNCHERS cache variable.  This cache variable initialization only occurs if CTEST_USE_LAUNCHERS is not already defined.

Added in version 3.8: If CTEST_USE_LAUNCHERS is on in a ctest -S script the ctest_configure command will add -DCTEST_USE_LAUNCHERS:BOOL=TRUE to the cmake command used to configure the project.

Functions to help assemble a standalone Qt4 executable.

A collection of CMake utility functions useful for deploying Qt4 executables.

The following functions are provided by this module:

write_qt4_conf
resolve_qt4_paths
fixup_qt4_executable
install_qt4_plugin_path
install_qt4_plugin
install_qt4_executable

Requires CMake 2.6 or greater because it uses function and PARENT_SCOPE.  Also depends on BundleUtilities.cmake.

write_qt4_conf(<qt_conf_dir> <qt_conf_contents>)

Writes a qt.conf file with the <qt_conf_contents> into <qt_conf_dir>.

resolve_qt4_paths(<paths_var> [<executable_path>])

Loop through <paths_var> list and if any don't exist resolve them relative to the <executable_path> (if supplied) or the CMAKE_INSTALL_PREFIX.

fixup_qt4_executable(<executable>
  [<qtplugins> <libs> <dirs> <plugins_dir> <request_qt_conf>])

Copies Qt plugins, writes a Qt configuration file (if needed) and fixes up a Qt4 executable using BundleUtilities so it is standalone and can be drag-and-drop copied to another machine as long as all of the system libraries are compatible.

<executable> should point to the executable to be fixed-up.

<qtplugins> should contain a list of the names or paths of any Qt plugins to be installed.

<libs> will be passed to BundleUtilities and should be a list of any already installed plugins, libraries or executables to also be fixed-up.

<dirs> will be passed to BundleUtilities and should contain and directories to be searched to find library dependencies.

<plugins_dir> allows an custom plugins directory to be used.

<request_qt_conf> will force a qt.conf file to be written even if not needed.

install_qt4_plugin_path(plugin executable copy installed_plugin_path_var
                        <plugins_dir> <component> <configurations>)

Install (or copy) a resolved <plugin> to the default plugins directory (or <plugins_dir>) relative to <executable> and store the result in <installed_plugin_path_var>.

If <copy> is set to TRUE then the plugins will be copied rather than installed.  This is to allow this module to be used at CMake time rather than install time.

If <component> is set then anything installed will use this COMPONENT.

install_qt4_plugin(plugin executable copy installed_plugin_path_var
                   <plugins_dir> <component>)

Install (or copy) an unresolved <plugin> to the default plugins directory (or <plugins_dir>) relative to <executable> and store the result in <installed_plugin_path_var>.  See documentation of INSTALL_QT4_PLUGIN_PATH.

install_qt4_executable(<executable>
  [<qtplugins> <libs> <dirs> <plugins_dir> <request_qt_conf> <component>])

Installs Qt plugins, writes a Qt configuration file (if needed) and fixes up a Qt4 executable using BundleUtilities so it is standalone and can be drag-and-drop copied to another machine as long as all of the system libraries are compatible.  The executable will be fixed-up at install time.  <component> is the COMPONENT used for bundle fixup and plugin installation.  See documentation of FIXUP_QT4_BUNDLE.

Manage data files stored outside source tree

Use this module to unambiguously reference data files stored outside the source tree and fetch them at build time from arbitrary local and remote content-addressed locations.  Functions provided by this module recognize arguments with the syntax DATA{<name>} as references to external data, replace them with full paths to local copies of those data, and create build rules to fetch and update the local copies.

For example:

include(ExternalData)
set(ExternalData_URL_TEMPLATES "file:///local/%(algo)/%(hash)"
                               "file:////host/share/%(algo)/%(hash)"
                               "http://data.org/%(algo)/%(hash)")
ExternalData_Add_Test(MyData
  NAME MyTest
  COMMAND MyExe DATA{MyInput.png}
  )
ExternalData_Add_Target(MyData)

When test MyTest runs the DATA{MyInput.png} argument will be replaced by the full path to a real instance of the data file MyInput.png on disk.  If the source tree contains a content link such as MyInput.png.md5 then the MyData target creates a real MyInput.png in the build tree.

ExternalData_Expand_Arguments

The ExternalData_Expand_Arguments function evaluates DATA{} references in its arguments and constructs a new list of arguments:

ExternalData_Expand_Arguments(
  <target>   # Name of data management target
  <outVar>   # Output variable
  [args...]  # Input arguments, DATA{} allowed
  )

It replaces each DATA{} reference in an argument with the full path of a real data file on disk that will exist after the <target> builds.

ExternalData_Add_Test

The ExternalData_Add_Test function wraps around the CMake add_test() command but supports DATA{} references in its arguments:

ExternalData_Add_Test(
  <target>   # Name of data management target
  ...        # Arguments of add_test(), DATA{} allowed
  )

It passes its arguments through ExternalData_Expand_Arguments and then invokes the add_test() command using the results.

Changed in version 3.31: If the arguments after <target> define a test with an executable that is a CMake target, empty values in the TEST_LAUNCHER and CROSSCOMPILING_EMULATOR properties of that target are preserved.  See policy CMP0178.

ExternalData_Add_Target

The ExternalData_Add_Target function creates a custom target to manage local instances of data files stored externally:

ExternalData_Add_Target(
  <target>                  # Name of data management target
  [SHOW_PROGRESS <ON|OFF>]  # Show progress during the download
  )

It creates custom commands in the target as necessary to make data files available for each DATA{} reference previously evaluated by other functions provided by this module. Data files may be fetched from one of the URL templates specified in the ExternalData_URL_TEMPLATES variable, or may be found locally in one of the paths specified in the ExternalData_OBJECT_STORES variable.

Added in version 3.20: The SHOW_PROGRESS argument may be passed to suppress progress information during the download of objects. If not provided, it defaults to OFF for Ninja and Ninja Multi-Config generators and ON otherwise.

Typically only one target is needed to manage all external data within a project.  Call this function once at the end of configuration after all data references have been processed.

The following variables configure behavior.  They should be set before calling any of the functions provided by this module.

ExternalData_BINARY_ROOT

The ExternalData_BINARY_ROOT variable may be set to the directory to hold the real data files named by expanded DATA{} references.  The default is CMAKE_BINARY_DIR.  The directory layout will mirror that of content links under ExternalData_SOURCE_ROOT.

ExternalData_CUSTOM_SCRIPT_<key>

Added in version 3.2.

Specify a full path to a .cmake custom fetch script identified by <key> in entries of the ExternalData_URL_TEMPLATES list. See Custom Fetch Scripts.

ExternalData_LINK_CONTENT

The ExternalData_LINK_CONTENT variable may be set to the name of a supported hash algorithm to enable automatic conversion of real data files referenced by the DATA{} syntax into content links.  For each such <file> a content link named <file><ext> is created.  The original file is renamed to the form .ExternalData_<algo>_<hash> to stage it for future transmission to one of the locations in the list of URL templates (by means outside the scope of this module).  The data fetch rule created for the content link will use the staged object if it cannot be found using any URL template.

ExternalData_NO_SYMLINKS

Added in version 3.3.

The real data files named by expanded DATA{} references may be made available under ExternalData_BINARY_ROOT using symbolic links on some platforms.  The ExternalData_NO_SYMLINKS variable may be set to disable use of symbolic links and enable use of copies instead.

ExternalData_OBJECT_STORES

The ExternalData_OBJECT_STORES variable may be set to a list of local directories that store objects using the layout <dir>/%(algo)/%(hash). These directories will be searched first for a needed object.  If the object is not available in any store then it will be fetched remotely using the URL templates and added to the first local store listed.  If no stores are specified the default is a location inside the build tree.

ExternalData_SERIES_PARSE

ExternalData_SERIES_PARSE_PREFIX

ExternalData_SERIES_PARSE_NUMBER

ExternalData_SERIES_PARSE_SUFFIX

ExternalData_SERIES_MATCH

See Referencing File Series.

ExternalData_SOURCE_ROOT

The ExternalData_SOURCE_ROOT variable may be set to the highest source directory containing any path named by a DATA{} reference.  The default is CMAKE_SOURCE_DIR.  ExternalData_SOURCE_ROOT and CMAKE_SOURCE_DIR must refer to directories within a single source distribution (e.g.  they come together in one tarball).

ExternalData_TIMEOUT_ABSOLUTE

The ExternalData_TIMEOUT_ABSOLUTE variable sets the download absolute timeout, in seconds, with a default of 300 seconds. Set to 0 to disable enforcement.

ExternalData_TIMEOUT_INACTIVITY

The ExternalData_TIMEOUT_INACTIVITY variable sets the download inactivity timeout, in seconds, with a default of 60 seconds. Set to 0 to disable enforcement.

ExternalData_URL_ALGO_<algo>_<key>

Added in version 3.3.

Specify a custom URL component to be substituted for URL template placeholders of the form %(algo:<key>), where <key> is a valid C identifier, when fetching an object referenced via hash algorithm <algo>.  If not defined, the default URL component is just <algo> for any <key>.

ExternalData_URL_TEMPLATES

The ExternalData_URL_TEMPLATES may be set to provide a list of URL templates using the placeholders %(algo) and %(hash) in each template.  Data fetch rules try each URL template in order by substituting the hash algorithm name for %(algo) and the hash value for %(hash).  Alternatively one may use %(algo:<key>) with ExternalData_URL_ALGO_<algo>_<key> variables to gain more flexibility in remote URLs.

The DATA{} syntax is literal and the <name> is a full or relative path within the source tree.  The source tree must contain either a real data file at <name> or a "content link" at <name><ext> containing a hash of the real file using a hash algorithm corresponding to <ext>. For example, the argument DATA{img.png} may be satisfied by either a real img.png file in the current source directory or a img.png.md5 file containing its MD5 sum.

Added in version 3.8: Multiple content links of the same name with different hash algorithms are supported (e.g. img.png.sha256 and img.png.sha1) so long as they all correspond to the same real file.  This allows objects to be fetched from sources indexed by different hash algorithms.

The DATA{} syntax can be told to fetch a file series using the form DATA{<name>,:}, where the : is literal.  If the source tree contains a group of files or content links named like a series then a reference to one member adds rules to fetch all of them.  Although all members of a series are fetched, only the file originally named by the DATA{} argument is substituted for it.  The default configuration recognizes file series names ending with #.ext, _#.ext, .#.ext, or -#.ext where # is a sequence of decimal digits and .ext is any single extension.  Configure it with a regex that parses <number> and <suffix> parts from the end of <name>:

ExternalData_SERIES_PARSE = regex of the form (<number>)(<suffix>)$

For more complicated cases set:

ExternalData_SERIES_PARSE = regex with at least two () groups
ExternalData_SERIES_PARSE_PREFIX = <prefix> regex group number, if any
ExternalData_SERIES_PARSE_NUMBER = <number> regex group number
ExternalData_SERIES_PARSE_SUFFIX = <suffix> regex group number

Configure series number matching with a regex that matches the <number> part of series members named <prefix><number><suffix>:

ExternalData_SERIES_MATCH = regex matching <number> in all series members

Note that the <suffix> of a series does not include a hash-algorithm extension.

The DATA{} syntax can alternatively match files associated with the named file and contained in the same directory.  Associated files may be specified by options using the syntax DATA{<name>,<opt1>,<opt2>,...}.  Each option may specify one file by name or specify a regular expression to match file names using the syntax REGEX:<regex>.  For example, the arguments:

DATA{MyData/MyInput.mhd,MyInput.img}                   # File pair
DATA{MyData/MyFrames00.png,REGEX:MyFrames[0-9]+\\.png} # Series

will pass MyInput.mha and MyFrames00.png on the command line but ensure that the associated files are present next to them.

The DATA{} syntax may reference a directory using a trailing slash and a list of associated files.  The form DATA{<name>/,<opt1>,<opt2>,...} adds rules to fetch any files in the directory that match one of the associated file options.  For example, the argument DATA{MyDataDir/,REGEX:.*} will pass the full path to a MyDataDir directory on the command line and ensure that the directory contains files corresponding to every file or content link in the MyDataDir source directory.

Added in version 3.3: In order to match associated files in subdirectories, specify a RECURSE: option, e.g. DATA{MyDataDir/,RECURSE:,REGEX:.*}.

The following hash algorithms are supported:

%(algo)     <ext>     Description
-------     -----     -----------
MD5         .md5      Message-Digest Algorithm 5, RFC 1321
SHA1        .sha1     US Secure Hash Algorithm 1, RFC 3174
SHA224      .sha224   US Secure Hash Algorithms, RFC 4634
SHA256      .sha256   US Secure Hash Algorithms, RFC 4634
SHA384      .sha384   US Secure Hash Algorithms, RFC 4634
SHA512      .sha512   US Secure Hash Algorithms, RFC 4634
SHA3_224    .sha3-224 Keccak SHA-3
SHA3_256    .sha3-256 Keccak SHA-3
SHA3_384    .sha3-384 Keccak SHA-3
SHA3_512    .sha3-512 Keccak SHA-3

Added in version 3.8: Added the SHA3_* hash algorithms.

Note that the hashes are used only for unique data identification and download verification.

Added in version 3.2.

When a data file must be fetched from one of the URL templates specified in the ExternalData_URL_TEMPLATES variable, it is normally downloaded using the file(DOWNLOAD) command. One may specify usage of a custom fetch script by using a URL template of the form ExternalDataCustomScript://<key>/<loc>. The <key> must be a C identifier, and the <loc> must contain the %(algo) and %(hash) placeholders. A variable corresponding to the key, ExternalData_CUSTOM_SCRIPT_<key>, must be set to the full path to a .cmake script file.  The script will be included to perform the actual fetch, and provided with the following variables:

ExternalData_CUSTOM_LOCATION

When a custom fetch script is loaded, this variable is set to the location part of the URL, which will contain the substituted hash algorithm name and content hash value.

ExternalData_CUSTOM_FILE

When a custom fetch script is loaded, this variable is set to the full path to a file in which the script must store the fetched content.  The name of the file is unspecified and should not be interpreted in any way.

The custom fetch script is expected to store fetched content in the file or set a variable:

ExternalData_CUSTOM_ERROR

When a custom fetch script fails to fetch the requested content, it must set this variable to a short one-line message describing the reason for failure.

ExternalProject_Add

The ExternalProject_Add() function creates a custom target to drive download, update/patch, configure, build, install and test steps of an external project:

ExternalProject_Add(<name> [<option>...])

The individual steps within the process can be driven independently if required (e.g. for CDash submission) and extra custom steps can be defined, along with the ability to control the step dependencies. The directory structure used for the management of the external project can also be customized. The function supports a large number of options which can be used to tailor the external project behavior.

Most of the time, the default directory layout is sufficient. It is largely an implementation detail that the main project usually doesn't need to change. In some circumstances, however, control over the directory layout can be useful or necessary. The directory options are potentially more useful from the point of view that the main build can use the ExternalProject_Get_Property() command to retrieve their values, thereby allowing the main project to refer to build artifacts of the external project.

PREFIX <dir>

Root directory for the external project. Unless otherwise noted below, all other directories associated with the external project will be created under here.

TMP_DIR <dir>

Directory in which to store temporary files.

STAMP_DIR <dir>

Directory in which to store the timestamps of each step. Log files from individual steps are also created in here unless overridden by LOG_DIR (see Logging Options below).

LOG_DIR <dir>

Added in version 3.14.

Directory in which to store the logs of each step.

DOWNLOAD_DIR <dir>

Directory in which to store downloaded files before unpacking them. This directory is only used by the URL download method, all other download methods use SOURCE_DIR directly instead.

SOURCE_DIR <dir>

Source directory into which downloaded contents will be unpacked, or for non-URL download methods, the directory in which the repository should be checked out, cloned, etc. If no download method is specified, this must point to an existing directory where the external project has already been unpacked or cloned/checked out.

NOTE:

If a download method is specified, any existing contents of the source directory may be deleted. Only the URL download method checks whether this directory is either missing or empty before initiating the download, stopping with an error if it is not empty. All other download methods silently discard any previous contents of the source directory.

BINARY_DIR <dir>

Specify the build directory location. This option is ignored if BUILD_IN_SOURCE is enabled.

INSTALL_DIR <dir>

Installation prefix to be placed in the <INSTALL_DIR> placeholder. This does not actually configure the external project to install to the given prefix. That must be done by passing appropriate arguments to the external project configuration step, e.g. using <INSTALL_DIR>.

If any of the above ..._DIR options are not specified, their defaults are computed as follows. If the PREFIX option is given or the EP_PREFIX directory property is set, then an external project is built and installed under the specified prefix:

TMP_DIR      = <prefix>/tmp
STAMP_DIR    = <prefix>/src/<name>-stamp
DOWNLOAD_DIR = <prefix>/src
SOURCE_DIR   = <prefix>/src/<name>
BINARY_DIR   = <prefix>/src/<name>-build
INSTALL_DIR  = <prefix>
LOG_DIR      = <STAMP_DIR>

Otherwise, if the EP_BASE directory property is set then components of an external project are stored under the specified base:

TMP_DIR      = <base>/tmp/<name>
STAMP_DIR    = <base>/Stamp/<name>
DOWNLOAD_DIR = <base>/Download/<name>
SOURCE_DIR   = <base>/Source/<name>
BINARY_DIR   = <base>/Build/<name>
INSTALL_DIR  = <base>/Install/<name>
LOG_DIR      = <STAMP_DIR>

If no PREFIX, EP_PREFIX, or EP_BASE is specified, then the default is to set PREFIX to <name>-prefix. Relative paths are interpreted with respect to CMAKE_CURRENT_BINARY_DIR at the point where ExternalProject_Add() is called.

A download method can be omitted if the SOURCE_DIR option is used to point to an existing non-empty directory. Otherwise, one of the download methods below must be specified (multiple download methods should not be given) or a custom DOWNLOAD_COMMAND provided.

DOWNLOAD_COMMAND <cmd>...

Overrides the command used for the download step (generator expressions are supported). If this option is specified, all other download options will be ignored. Providing an empty string for <cmd> effectively disables the download step.

URL <url1> [<url2>...]

List of paths and/or URL(s) of the external project's source. When more than one URL is given, they are tried in turn until one succeeds. A URL may be an ordinary path in the local file system (in which case it must be the only URL provided) or any downloadable URL supported by the file(DOWNLOAD) command. A local filesystem path may refer to either an existing directory or to an archive file, whereas a URL is expected to point to a file which can be treated as an archive. When an archive is used, it will be unpacked automatically unless the DOWNLOAD_NO_EXTRACT option is set to prevent it. The archive type is determined by inspecting the actual content rather than using logic based on the file extension.

Changed in version 3.7: Multiple URLs are allowed.

URL_HASH <algo>=<hashValue>

Hash of the archive file to be downloaded. The argument should be of the form <algo>=<hashValue> where algo can be any of the hashing algorithms supported by the file() command. Specifying this option is strongly recommended for URL downloads, as it ensures the integrity of the downloaded content. It is also used as a check for a previously downloaded file, allowing connection to the remote location to be avoided altogether if the local directory already has a file from an earlier download that matches the specified hash.

URL_MD5 <md5>

Equivalent to URL_HASH MD5=<md5>.

DOWNLOAD_NAME <fname>

File name to use for the downloaded file. If not given, the end of the URL is used to determine the file name. This option is rarely needed, the default name is generally suitable and is not normally used outside of code internal to the ExternalProject module.

DOWNLOAD_EXTRACT_TIMESTAMP <bool>

Added in version 3.24.

When specified with a true value, the timestamps of the extracted files will match those in the archive. When false, the timestamps of the extracted files will reflect the time at which the extraction was performed. If the download URL changes, timestamps based off those in the archive can result in dependent targets not being rebuilt when they potentially should have been. Therefore, unless the file timestamps are significant to the project in some way, use a false value for this option. If DOWNLOAD_EXTRACT_TIMESTAMP is not given, the default is false. See policy CMP0135.

DOWNLOAD_NO_EXTRACT <bool>

Added in version 3.6.

Allows the extraction part of the download step to be disabled by passing a boolean true value for this option. If this option is not given, the downloaded contents will be unpacked automatically if required. If extraction has been disabled, the full path to the downloaded file is available as <DOWNLOADED_FILE> in subsequent steps or as the property DOWNLOADED_FILE with the ExternalProject_Get_Property() command.

DOWNLOAD_NO_PROGRESS <bool>

Can be used to disable logging the download progress. If this option is not given, download progress messages will be logged.

TIMEOUT <seconds>

Maximum time allowed for file download operations.

INACTIVITY_TIMEOUT <seconds>

Added in version 3.19.

Terminate the operation after a period of inactivity.

HTTP_USERNAME <username>

Added in version 3.7.

Username for the download operation if authentication is required.

HTTP_PASSWORD <password>

Added in version 3.7.

Password for the download operation if authentication is required.

HTTP_HEADER <header1> [<header2>...]

Added in version 3.7.

Provides an arbitrary list of HTTP headers for the download operation. This can be useful for accessing content in systems like AWS, etc.

TLS_VERSION <min>

Added in version 3.30.

Specify minimum TLS version for https:// URLs.  If this option is not provided, the value of the CMAKE_TLS_VERSION variable or the CMAKE_TLS_VERSION environment variable will be used instead (see file(DOWNLOAD)).

This option also applies to git clone invocations, although the default behavior is different.  If none of the TLS_VERSION option, CMAKE_TLS_VERSION variable, or CMAKE_TLS_VERSION environment variable is specified, the behavior will be determined by git's default or a http.sslVersion git config option the user may have set at a global level.

TLS_VERIFY <bool>

Specifies whether certificate verification should be performed for https:// URLs.  If this option is not provided, the value of the CMAKE_TLS_VERIFY variable or the CMAKE_TLS_VERIFY environment variable will be used instead (see file(DOWNLOAD)). If neither of those is set, certificate verification will not be performed. In situations where URL_HASH cannot be provided, this option can be an alternative verification measure.

This option also applies to git clone invocations, although the default behavior is different.  If none of the TLS_VERIFY option, CMAKE_TLS_VERIFY variable, or CMAKE_TLS_VERIFY environment variable is specified, the behavior will be determined by git's default (true) or a http.sslVerify git config option the user may have set at a global level.

Changed in version 3.6: Previously this option did not apply to git clone invocations.

Changed in version 3.30: Previously the CMAKE_TLS_VERIFY environment variable was not checked.

TLS_CAINFO <file>

Specify a custom certificate authority file to use if TLS_VERIFY is enabled. If this option is not specified, the value of the CMAKE_TLS_CAINFO variable will be used instead (see file(DOWNLOAD))

NETRC <level>

Added in version 3.11.

Specify whether the .netrc file is to be used for operation. If this option is not specified, the value of the CMAKE_NETRC variable will be used instead (see file(DOWNLOAD)).  Valid levels are:

IGNORED

The .netrc file is ignored. This is the default.

OPTIONAL

The .netrc file is optional, and information in the URL is preferred.  The file will be scanned to find which ever information is not specified in the URL.

REQUIRED

The .netrc file is required, and information in the URL is ignored.

NETRC_FILE <file>

Added in version 3.11.

Specify an alternative .netrc file to the one in your home directory if the NETRC level is OPTIONAL or REQUIRED. If this option is not specified, the value of the CMAKE_NETRC_FILE variable will be used instead (see file(DOWNLOAD))

Added in version 3.1: Added support for tbz2, .tar.xz, .txz, and .7z extensions.

NOTE: A git version of 1.6.5 or later is required if this download method is used.

GIT_REPOSITORY <url>

URL of the git repository. Any URL understood by the git command may be used.

Changed in version 3.27: A relative URL will be resolved based on the parent project's remote, subject to CMP0150.  See the policy documentation for how the remote is selected, including conditions where the remote selection can fail.  Local filesystem remotes should always use absolute paths.

GIT_TAG <tag>

Git branch name, tag or commit hash. Note that branch names and tags should generally be specified as remote names (i.e. origin/myBranch rather than simply myBranch). This ensures that if the remote end has its tag moved or branch rebased or history rewritten, the local clone will still be updated correctly. In general, however, specifying a commit hash should be preferred for a number of reasons:

• If the local clone already has the commit corresponding to the hash, no git fetch needs to be performed to check for changes each time CMake is re-run. This can result in a significant speed up if many external projects are being used.
• Using a specific git hash ensures that the main project's own history is fully traceable to a specific point in the external project's evolution. If a branch or tag name is used instead, then checking out a specific commit of the main project doesn't necessarily pin the whole build to a specific point in the life of the external project. The lack of such deterministic behavior makes the main project lose traceability and repeatability.

If GIT_SHALLOW is enabled then GIT_TAG works only with branch names and tags.  A commit hash is not allowed.

Note that if not provided, GIT_TAG defaults to master, not the default Git branch name.

GIT_REMOTE_NAME <name>

The optional name of the remote. If this option is not specified, it defaults to origin.

GIT_SUBMODULES <module>...

Specific git submodules that should also be updated. If this option is not provided, all git submodules will be updated.

Changed in version 3.16: When CMP0097 is set to NEW, if this value is set to an empty string then no submodules are initialized or updated.

GIT_SUBMODULES_RECURSE <bool>

Added in version 3.17.

Specify whether git submodules (if any) should update recursively by passing the --recursive flag to git submodule update. If not specified, the default is on.

GIT_SHALLOW <bool>

Added in version 3.6.

When this option is enabled, the git clone operation will be given the --depth 1 option. This performs a shallow clone, which avoids downloading the whole history and instead retrieves just the commit denoted by the GIT_TAG option.

GIT_PROGRESS <bool>

Added in version 3.8.

When enabled, this option instructs the git clone operation to report its progress by passing it the --progress option. Without this option, the clone step for large projects may appear to make the build stall, since nothing will be logged until the clone operation finishes. While this option can be used to provide progress to prevent the appearance of the build having stalled, it may also make the build overly noisy if lots of external projects are used.

GIT_CONFIG <option1> [<option2>...]

Added in version 3.8.

Specify a list of config options to pass to git clone. Each option listed will be transformed into its own --config <option> on the git clone command line, with each option required to be in the form key=value.

GIT_REMOTE_UPDATE_STRATEGY <strategy>

Added in version 3.18.

When GIT_TAG refers to a remote branch, this option can be used to specify how the update step behaves.  The <strategy> must be one of the following:

CHECKOUT

Ignore the local branch and always checkout the branch specified by GIT_TAG.

REBASE

Try to rebase the current branch to the one specified by GIT_TAG. If there are local uncommitted changes, they will be stashed first and popped again after rebasing.  If rebasing or popping stashed changes fail, abort the rebase and halt with an error. When GIT_REMOTE_UPDATE_STRATEGY is not present, this is the default strategy unless the default has been overridden with CMAKE_EP_GIT_REMOTE_UPDATE_STRATEGY (see below). Note that if the branch specified in GIT_TAG is different to the upstream branch currently being tracked, it is not safe to perform a rebase. In that situation, REBASE will silently be treated as CHECKOUT instead.

REBASE_CHECKOUT

Same as REBASE except if the rebase fails, an annotated tag will be created at the original HEAD position from before the rebase and then checkout GIT_TAG just like the CHECKOUT strategy. The message stored on the annotated tag will give information about what was attempted and the tag name will include a timestamp so that each failed run will add a new tag.  This strategy ensures no changes will be lost, but updates should always succeed if GIT_TAG refers to a valid ref unless there are uncommitted changes that cannot be popped successfully.

The variable CMAKE_EP_GIT_REMOTE_UPDATE_STRATEGY can be set to override the default strategy.  This variable should not be set by a project, it is intended for the user to set.  It is primarily intended for use in continuous integration scripts to ensure that when history is rewritten on a remote branch, the build doesn't end up with unintended changes or failed builds resulting from conflicts during rebase operations.

SVN_REPOSITORY <url>

URL of the Subversion repository.

SVN_REVISION -r<rev>

Revision to checkout from the Subversion repository.

SVN_USERNAME <username>

Username for the Subversion checkout and update.

SVN_PASSWORD <password>

Password for the Subversion checkout and update.

SVN_TRUST_CERT <bool>

Specifies whether to trust the Subversion server site certificate. If enabled, the --trust-server-cert option is passed to the svn checkout and update commands.

HG_REPOSITORY <url>

URL of the mercurial repository.

HG_TAG <tag>

Mercurial branch name, tag or commit id.

CVS_REPOSITORY <cvsroot>

CVSROOT of the CVS repository.

CVS_MODULE <mod>

Module to checkout from the CVS repository.

CVS_TAG <tag>

Tag to checkout from the CVS repository.

Whenever CMake is re-run, by default the external project's sources will be updated if the download method supports updates (e.g. a git repository would be checked if the GIT_TAG does not refer to a specific commit).

UPDATE_COMMAND <cmd>...

Overrides the download method's update step with a custom command. The command may use generator expressions.

UPDATE_DISCONNECTED <bool>

Added in version 3.2.

When enabled, this option causes the update step to be skipped (but see below for changed behavior where this is not the case). It does not prevent the download step. The update step can still be added as a step target (see ExternalProject_Add_StepTargets()) and called manually. This is useful if you want to allow developers to build the project when disconnected from the network (the network may still be needed for the download step though).

Changed in version 3.27: When UPDATE_DISCONNECTED is true, the update step will be executed if any details about the update or download step are changed. Furthermore, if using the git download/update method, the update logic will be modified to skip attempts to contact the remote. If the GIT_TAG mentions a ref that is not known locally, the update step will halt with a fatal error.

When this option is present, it is generally advisable to make the value a cache variable under the developer's control rather than hard-coding it. If this option is not present, the default value is taken from the EP_UPDATE_DISCONNECTED directory property. If that is also not defined, updates are performed as normal. The EP_UPDATE_DISCONNECTED directory property is intended as a convenience for controlling the UPDATE_DISCONNECTED behavior for an entire section of a project's directory hierarchy and may be a more convenient method of giving developers control over whether or not to perform updates (assuming the project also provides a cache variable or some other convenient method for setting the directory property).

This may cause a step target to be created automatically for the download step.  See policy CMP0114.

PATCH_COMMAND <cmd>...

Specifies a custom command to patch the sources after an update. By default, no patch command is defined. Note that it can be quite difficult to define an appropriate patch command that performs robustly, especially for download methods such as git where changing the GIT_TAG will not discard changes from a previous patch, but the patch command will be called again after updating to the new tag.

The configure step is run after the download and update steps. By default, the external project is assumed to be a CMake project, but this can be overridden if required.

CONFIGURE_COMMAND <cmd>...

The default configure command runs CMake with a few options based on the main project.  The options added are typically only those needed to use the same generator as the main project, but the CMAKE_GENERATOR option can be given to override this.  The project is responsible for adding any toolchain details, flags or other settings it wants to reuse from the main project or otherwise specify (see CMAKE_ARGS, CMAKE_CACHE_ARGS and CMAKE_CACHE_DEFAULT_ARGS below).

For non-CMake external projects, the CONFIGURE_COMMAND option must be used to override the default configure command (generator expressions are supported). For projects that require no configure step, specify this option with an empty string as the command to execute.

CMAKE_COMMAND /.../cmake

Specify an alternative cmake executable for the configure step (use an absolute path). This is generally not recommended, since it is usually desirable to use the same CMake version throughout the whole build. This option is ignored if a custom configure command has been specified with CONFIGURE_COMMAND.

CMAKE_GENERATOR <gen>

Override the CMake generator used for the configure step. Without this option, the same generator as the main build will be used. This option is ignored if a custom configure command has been specified with the CONFIGURE_COMMAND option.

CMAKE_GENERATOR_PLATFORM <platform>

Added in version 3.1.

Pass a generator-specific platform name to the CMake command (see CMAKE_GENERATOR_PLATFORM). It is an error to provide this option without the CMAKE_GENERATOR option.

CMAKE_GENERATOR_TOOLSET <toolset>

Pass a generator-specific toolset name to the CMake command (see CMAKE_GENERATOR_TOOLSET). It is an error to provide this option without the CMAKE_GENERATOR option.

CMAKE_GENERATOR_INSTANCE <instance>

Added in version 3.11.

Pass a generator-specific instance selection to the CMake command (see CMAKE_GENERATOR_INSTANCE). It is an error to provide this option without the CMAKE_GENERATOR option.

CMAKE_ARGS <arg>...

The specified arguments are passed to the cmake command line. They can be any argument the cmake command understands, not just cache values defined by -D... arguments (see also CMake Options).

Added in version 3.3: Arguments may use generator expressions.

CMAKE_CACHE_ARGS <arg>...

This is an alternate way of specifying cache variables where command line length issues may become a problem. The arguments are expected to be in the form -Dvar:STRING=value, which are then transformed into CMake set() commands with the FORCE option used. These set() commands are written to a pre-load script which is then applied using the cmake -C command line option.

Added in version 3.3: Arguments may use generator expressions.

CMAKE_CACHE_DEFAULT_ARGS <arg>...

Added in version 3.2.

This is the same as the CMAKE_CACHE_ARGS option except the set() commands do not include the FORCE keyword. This means the values act as initial defaults only and will not override any variables already set from a previous run. Use this option with care, as it can lead to different behavior depending on whether the build starts from a fresh build directory or reuses previous build contents.

Added in version 3.15: If the CMake generator is the Green Hills MULTI and not overridden, the original project's settings for the GHS toolset and target system customization cache variables are propagated into the external project.

SOURCE_SUBDIR <dir>

Added in version 3.7.

When no CONFIGURE_COMMAND option is specified, the configure step assumes the external project has a CMakeLists.txt file at the top of its source tree (i.e. in SOURCE_DIR). The SOURCE_SUBDIR option can be used to point to an alternative directory within the source tree to use as the top of the CMake source tree instead. This must be a relative path and it will be interpreted as being relative to SOURCE_DIR.

Added in version 3.14: When BUILD_IN_SOURCE option is enabled, the BUILD_COMMAND is used to point to an alternative directory within the source tree.

CONFIGURE_HANDLED_BY_BUILD <bool>

Added in version 3.20.

Enabling this option relaxes the dependencies of the configure step on other external projects to order-only. This means the configure step will be executed after its external project dependencies are built but it will not be marked dirty when one of its external project dependencies is rebuilt. This option can be enabled when the build step is smart enough to figure out if the configure step needs to be rerun. CMake and Meson are examples of build systems whose build step is smart enough to know if the configure step needs to be rerun.

If the configure step assumed the external project uses CMake as its build system, the build step will also. Otherwise, the build step will assume a Makefile-based build and simply run make with no arguments as the default build step. This can be overridden with custom build commands if required.

If both the main project and the external project use make as their build tool, the build step of the external project is invoked as a recursive make using $(MAKE).  This will communicate some build tool settings from the main project to the external project.  If either the main project or external project is not using make, no build tool settings will be passed to the external project other than those established by the configure step (i.e. running ninja -v in the main project will not pass -v to the external project's build step, even if it also uses ninja as its build tool).

BUILD_COMMAND <cmd>...

Overrides the default build command (generator expressions are supported). If this option is not given, the default build command will be chosen to integrate with the main build in the most appropriate way (e.g. using recursive make for Makefile generators or cmake --build if the project uses a CMake build). This option can be specified with an empty string as the command to make the build step do nothing.

BUILD_IN_SOURCE <bool>

When this option is enabled, the build will be done directly within the external project's source tree. This should generally be avoided, the use of a separate build directory is usually preferred, but it can be useful when the external project assumes an in-source build. The BINARY_DIR option should not be specified if building in-source.

BUILD_ALWAYS <bool>

Enabling this option forces the build step to always be run. This can be the easiest way to robustly ensure that the external project's own build dependencies are evaluated rather than relying on the default success timestamp-based method. This option is not normally needed unless developers are expected to modify something the external project's build depends on in a way that is not detectable via the step target dependencies (e.g. SOURCE_DIR is used without a download method and developers might modify the sources in SOURCE_DIR).

BUILD_BYPRODUCTS <file>...

Added in version 3.2.

Specifies files that will be generated by the build command but which might or might not have their modification time updated by subsequent builds. This may also be required to explicitly declare dependencies when using the Ninja generator. These ultimately get passed through as BYPRODUCTS to the build step's own underlying call to add_custom_command(), which has additional documentation.

BUILD_JOB_SERVER_AWARE <bool>

Added in version 3.28.

Specifies that the build step is aware of the GNU Make job server. See the add_custom_command() documentation of its JOB_SERVER_AWARE option for details.  This option is relevant only when an explicit BUILD_COMMAND is specified.

If the configure step assumed the external project uses CMake as its build system, the install step will also. Otherwise, the install step will assume a Makefile-based build and simply run make install as the default build step. This can be overridden with custom install commands if required.

INSTALL_COMMAND <cmd>...

The external project's own install step is invoked as part of the main project's build. It is done after the external project's build step and may be before or after the external project's test step (see the TEST_BEFORE_INSTALL option below). The external project's install rules are not part of the main project's install rules, so if anything from the external project should be installed as part of the main build, these need to be specified in the main build as additional install() commands. The default install step builds the install target of the external project, but this can be overridden with a custom command using this option (generator expressions are supported). Passing an empty string as the <cmd> makes the install step do nothing.

INSTALL_BYPRODUCTS <file>...

Added in version 3.26.

Specifies files that will be generated by the install command but which might or might not have their modification time updated by subsequent installs. This may also be required to explicitly declare dependencies when using the Ninja generator. These ultimately get passed through as BYPRODUCTS to the install step's own underlying call to add_custom_command(), which has additional documentation.

NOTE:

If the CMAKE_INSTALL_MODE environment variable is set when the main project is built, it will only have an effect if the following conditions are met:

• The main project's configure step assumed the external project uses CMake as its build system.
• The external project's install command actually runs. Note that due to the way ExternalProject may use timestamps internally, if nothing the install step depends on needs to be re-executed, the install command might also not need to run.

Note also that ExternalProject does not check whether the CMAKE_INSTALL_MODE environment variable changes from one run to another.

The test step is only defined if at least one of the following TEST_... options are provided.

TEST_COMMAND <cmd>...

Overrides the default test command (generator expressions are supported). If this option is not given, the default behavior of the test step is to build the external project's own test target. This option can be specified with <cmd> as an empty string, which allows the test step to still be defined, but it will do nothing. Do not specify any of the other TEST_... options if providing an empty string as the test command, but prefer to omit all TEST_... options altogether if the test step target is not needed.

TEST_BEFORE_INSTALL <bool>

When this option is enabled, the test step will be executed before the install step. The default behavior is for the test step to run after the install step.

TEST_AFTER_INSTALL <bool>

This option is mainly useful as a way to indicate that the test step is desired but all default behavior is sufficient. Specifying this option with a boolean true value ensures the test step is defined and that it comes after the install step. If both TEST_BEFORE_INSTALL and TEST_AFTER_INSTALL are enabled, the latter is silently ignored.

TEST_EXCLUDE_FROM_MAIN <bool>

Added in version 3.2.

If enabled, the main build's default ALL target will not depend on the test step. This can be a useful way of ensuring the test step is defined but only gets invoked when manually requested. This may cause a step target to be created automatically for either the install or build step.  See policy CMP0114.

Each of the following LOG_... options can be used to wrap the relevant step in a script to capture its output to files. The log files will be created in LOG_DIR if supplied or otherwise the STAMP_DIR directory with step-specific file names.

LOG_DOWNLOAD <bool>

When enabled, the output of the download step is logged to files.

LOG_UPDATE <bool>

When enabled, the output of the update step is logged to files.

LOG_PATCH <bool>

Added in version 3.14.

When enabled, the output of the patch step is logged to files.

LOG_CONFIGURE <bool>

When enabled, the output of the configure step is logged to files.

LOG_BUILD <bool>

When enabled, the output of the build step is logged to files.

LOG_INSTALL <bool>

When enabled, the output of the install step is logged to files.

LOG_TEST <bool>

When enabled, the output of the test step is logged to files.

LOG_MERGED_STDOUTERR <bool>

Added in version 3.14.

When enabled, stdout and stderr will be merged for any step whose output is being logged to files.

LOG_OUTPUT_ON_FAILURE <bool>

Added in version 3.14.

This option only has an effect if at least one of the other LOG_<step> options is enabled.  If an error occurs for a step which has logging to file enabled, that step's output will be printed to the console if LOG_OUTPUT_ON_FAILURE is set to true.  For cases where a large amount of output is recorded, just the end of that output may be printed to the console.

Added in version 3.4.

Steps can be given direct access to the terminal in some cases. Giving a step access to the terminal may allow it to receive terminal input if required, such as for authentication details not provided by other options. With the Ninja generator, these options place the steps in the console job pool. Each step can be given access to the terminal individually via the following options:

USES_TERMINAL_DOWNLOAD <bool>

Give the download step access to the terminal.

USES_TERMINAL_UPDATE <bool>

Give the update step access to the terminal.

USES_TERMINAL_PATCH <bool>

Added in version 3.23.

Give the patch step access to the terminal.

USES_TERMINAL_CONFIGURE <bool>

Give the configure step access to the terminal.

USES_TERMINAL_BUILD <bool>

Give the build step access to the terminal.

USES_TERMINAL_INSTALL <bool>

Give the install step access to the terminal.

USES_TERMINAL_TEST <bool>

Give the test step access to the terminal.

DEPENDS <targets>...

Specify other targets on which the external project depends. The other targets will be brought up to date before any of the external project's steps are executed. Because the external project uses additional custom targets internally for each step, the DEPENDS option is the most convenient way to ensure all of those steps depend on the other targets. Simply doing add_dependencies(<name> <targets>) will not make any of the steps dependent on <targets>.

EXCLUDE_FROM_ALL <bool>

When enabled, this option excludes the external project from the default ALL target of the main build.

STEP_TARGETS <step-target>...

Generate custom targets for the specified steps. This is required if the steps need to be triggered manually or if they need to be used as dependencies of other targets. If this option is not specified, the default value is taken from the EP_STEP_TARGETS directory property. See ExternalProject_Add_StepTargets() below for further discussion of the effects of this option.

INDEPENDENT_STEP_TARGETS <step-target>...

Deprecated since version 3.19: This is allowed only if policy CMP0114 is not set to NEW.

Generates custom targets for the specified steps and prevent these targets from having the usual dependencies applied to them. If this option is not specified, the default value is taken from the EP_INDEPENDENT_STEP_TARGETS directory property. This option is mostly useful for allowing individual steps to be driven independently, such as for a CDash setup where each step should be initiated and reported individually rather than as one whole build. See ExternalProject_Add_StepTargets() below for further discussion of the effects of this option.

LIST_SEPARATOR <sep>

For any of the various ..._COMMAND options, and CMAKE_ARGS, ExternalProject will replace <sep> with ; in the specified command lines. This can be used to ensure a command has a literal ; in it where direct usage would otherwise be interpreted as argument separators to CMake APIs instead. Note that the separator should be chosen to avoid being confused for non-list-separator usages of the sequence. For example, using LIST_SEPARATOR allows for passing list values to CMake cache variables on the command line:

ExternalProject_Add(example
  ... # Download options, etc.
  LIST_SEPARATOR ","
  CMAKE_ARGS "-DCMAKE_PREFIX_PATH:STRING=${first_prefix},${second_prefix}"
)

COMMAND <cmd>...

Any of the other ..._COMMAND options can have additional commands appended to them by following them with as many COMMAND ... options as needed (generator expressions are supported). For example:

ExternalProject_Add(example
  ... # Download options, etc.
  BUILD_COMMAND ${CMAKE_COMMAND} -E echo "Starting $<CONFIG> build"
  COMMAND       ${CMAKE_COMMAND} --build <BINARY_DIR> --config $<CONFIG>
  COMMAND       ${CMAKE_COMMAND} -E echo "$<CONFIG> build complete"
)

It should also be noted that each build step is created via a call to ExternalProject_Add_Step(). See that command's documentation for the automatic substitutions that are supported for some options.

ExternalProject_Get_Property

The ExternalProject_Get_Property() function retrieves external project target properties:

ExternalProject_Get_Property(<name> <prop1> [<prop2>...])

The function stores property values in variables of the same name. Property names correspond to the keyword argument names of ExternalProject_Add(). For example, the source directory might be retrieved like so:

ExternalProject_Get_property(myExtProj SOURCE_DIR)
message("Source dir of myExtProj = ${SOURCE_DIR}")

The ExternalProject_Add() function on its own is often sufficient for incorporating an external project into the main build. Certain scenarios require additional work to implement desired behavior, such as adding in a custom step or making steps available as manually triggerable targets. The ExternalProject_Add_Step(), ExternalProject_Add_StepTargets() and ExternalProject_Add_StepDependencies functions provide the lower level control needed to implement such step-level capabilities.

ExternalProject_Add_Step

The ExternalProject_Add_Step() function specifies an additional custom step for an external project defined by an earlier call to ExternalProject_Add():

ExternalProject_Add_Step(<name> <step> [<option>...])

<name> is the same as the name passed to the original call to ExternalProject_Add(). The specified <step> must not be one of the pre-defined steps (mkdir, download, update, patch, configure, build, install or test). The supported options are:

COMMAND <cmd>...

The command line to be executed by this custom step (generator expressions are supported). This option can be repeated multiple times to specify multiple commands to be executed in order.

COMMENT "<text>..."

Text to be printed when the custom step executes.

DEPENDEES <step>...

Other steps (custom or pre-defined) on which this step depends.

DEPENDERS <step>...

Other steps (custom or pre-defined) that depend on this new custom step.

DEPENDS <file>...

Files on which this custom step depends.

INDEPENDENT <bool>

Added in version 3.19.

Specifies whether this step is independent of the external dependencies specified by the ExternalProject_Add()'s DEPENDS option. The default is FALSE.  Steps marked as independent may depend only on other steps marked independent.  See policy CMP0114.

Note that this use of the term "independent" refers only to independence from external targets specified by the DEPENDS option and is orthogonal to a step's dependencies on other steps.

If a step target is created for an independent step by the ExternalProject_Add() STEP_TARGETS option or by the ExternalProject_Add_StepTargets() function, it will not depend on the external targets, but may depend on targets for other steps.

BYPRODUCTS <file>...

Added in version 3.2.

Files that will be generated by this custom step but which might or might not have their modification time updated by subsequent builds. This may also be required to explicitly declare dependencies when using the Ninja generator. This list of files will ultimately be passed through as the BYPRODUCTS option to the add_custom_command() used to implement the custom step internally, which has additional documentation.

ALWAYS <bool>

When enabled, this option specifies that the custom step should always be run (i.e. that it is always considered out of date).

JOB_SERVER_AWARE <bool>

Added in version 3.28.

Specifies that the custom step is aware of the GNU Make job server. See the add_custom_command() documentation of its JOB_SERVER_AWARE option for details.

EXCLUDE_FROM_MAIN <bool>

When enabled, this option specifies that the external project's main target does not depend on the custom step. This may cause step targets to be created automatically for the steps on which this step depends.  See policy CMP0114.

WORKING_DIRECTORY <dir>

Specifies the working directory to set before running the custom step's command. If this option is not specified, the directory will be the value of the CMAKE_CURRENT_BINARY_DIR at the point where ExternalProject_Add_Step() was called.

LOG <bool>

If set, this causes the output from the custom step to be captured to files in the external project's LOG_DIR if supplied or STAMP_DIR.

USES_TERMINAL <bool>

If enabled, this gives the custom step direct access to the terminal if possible.

The command line, comment, working directory and byproducts of every standard and custom step are processed to replace the tokens <SOURCE_DIR>, <SOURCE_SUBDIR>, <BINARY_DIR>, <INSTALL_DIR> <TMP_DIR>, <DOWNLOAD_DIR> and <DOWNLOADED_FILE> with their corresponding property values defined in the original call to ExternalProject_Add().

Added in version 3.3: Token replacement is extended to byproducts.

Added in version 3.11: The <DOWNLOAD_DIR> substitution token.

ExternalProject_Add_StepTargets

The ExternalProject_Add_StepTargets() function generates targets for the steps listed. The name of each created target will be of the form <name>-<step>:

ExternalProject_Add_StepTargets(<name> <step1> [<step2>...])

Creating a target for a step allows it to be used as a dependency of another target or to be triggered manually. Having targets for specific steps also allows them to be driven independently of each other by specifying targets on build command lines. For example, you may be submitting to a sub-project based dashboard where you want to drive the configure portion of the build, then submit to the dashboard, followed by the build portion, followed by tests. If you invoke a custom target that depends on a step halfway through the step dependency chain, then all the previous steps will also run to ensure everything is up to date.

Internally, ExternalProject_Add() calls ExternalProject_Add_Step() to create each step. If any STEP_TARGETS were specified, then ExternalProject_Add_StepTargets() will also be called after ExternalProject_Add_Step().  Even if a step is not mentioned in the STEP_TARGETS option, ExternalProject_Add_StepTargets() can still be called later to manually define a target for the step.

The STEP_TARGETS option for ExternalProject_Add() is generally the easiest way to ensure targets are created for specific steps of interest. For custom steps, ExternalProject_Add_StepTargets() must be called explicitly if a target should also be created for that custom step. An alternative to these two options is to populate the EP_STEP_TARGETS directory property.  It acts as a default for the step target options and can save having to repeatedly specify the same set of step targets when multiple external projects are being defined.

Added in version 3.19: If CMP0114 is set to NEW, step targets are fully responsible for holding the custom commands implementing their steps.  The primary target created by ExternalProject_Add depends on the step targets, and the step targets depend on each other.  The target-level dependencies match the file-level dependencies used by the custom commands for each step. The targets for steps created with ExternalProject_Add_Step()'s INDEPENDENT option do not depend on the external targets specified by ExternalProject_Add()'s DEPENDS option.  The predefined steps mkdir, download, update, and patch are independent.

If CMP0114 is not NEW, the following deprecated behavior is available:

• A deprecated NO_DEPENDS option may be specified immediately after the <name> and before the first step. If the NO_DEPENDS option is specified, the step target will not depend on the dependencies of the external project (i.e. on any dependencies of the <name> custom target created by ExternalProject_Add()). This is usually safe for the download, update and patch steps, since they do not typically require that the dependencies are updated and built. Using NO_DEPENDS for any of the other pre-defined steps, however, may break parallel builds. Only use NO_DEPENDS where it is certain that the named steps genuinely do not have dependencies. For custom steps, consider whether or not the custom commands require the dependencies to be configured, built and installed.
• The INDEPENDENT_STEP_TARGETS option for ExternalProject_Add(), or the EP_INDEPENDENT_STEP_TARGETS directory property, tells the function to call ExternalProject_Add_StepTargets() internally using the NO_DEPENDS option for the specified steps.

ExternalProject_Add_StepDependencies

Added in version 3.2.

The ExternalProject_Add_StepDependencies() function can be used to add dependencies to a step. The dependencies added must be targets CMake already knows about (these can be ordinary executable or library targets, custom targets or even step targets of another external project):

ExternalProject_Add_StepDependencies(<name> <step> <target1> [<target2>...])

This function takes care to set both target and file level dependencies and will ensure that parallel builds will not break. It should be used instead of add_dependencies() whenever adding a dependency for some of the step targets generated by the ExternalProject module.

The following example shows how to download and build a hypothetical project called FooBar from github:

include(ExternalProject)
ExternalProject_Add(foobar
  GIT_REPOSITORY    git@github.com:FooCo/FooBar.git
  GIT_TAG           origin/release/1.2.3
)

For the sake of the example, also define a second hypothetical external project called SecretSauce, which is downloaded from a web server. Two URLs are given to take advantage of a faster internal network if available, with a fallback to a slower external server. The project is a typical Makefile project with no configure step, so some of the default commands are overridden. The build is only required to build the sauce target:

find_program(MAKE_EXE NAMES gmake nmake make)
ExternalProject_Add(secretsauce
  URL               http://intranet.somecompany.com/artifacts/sauce-2.7.tgz
                    https://www.somecompany.com/downloads/sauce-2.7.zip
  URL_HASH          MD5=d41d8cd98f00b204e9800998ecf8427e
  CONFIGURE_COMMAND ""
  BUILD_COMMAND     ${MAKE_EXE} sauce
)

Suppose the build step of secretsauce requires that foobar must already be built. This could be enforced like so:

ExternalProject_Add_StepDependencies(secretsauce build foobar)

Another alternative would be to create a custom target for foobar's build step and make secretsauce depend on that rather than the whole foobar project. This would mean foobar only needs to be built, it doesn't need to run its install or test steps before secretsauce can be built. The dependency can also be defined along with the secretsauce project:

ExternalProject_Add_StepTargets(foobar build)
ExternalProject_Add(secretsauce
  URL               http://intranet.somecompany.com/artifacts/sauce-2.7.tgz
                    https://www.somecompany.com/downloads/sauce-2.7.zip
  URL_HASH          MD5=d41d8cd98f00b204e9800998ecf8427e
  CONFIGURE_COMMAND ""
  BUILD_COMMAND     ${MAKE_EXE} sauce
  DEPENDS           foobar-build
)

Instead of calling ExternalProject_Add_StepTargets(), the target could be defined along with the foobar project itself:

ExternalProject_Add(foobar
  GIT_REPOSITORY git@github.com:FooCo/FooBar.git
  GIT_TAG        origin/release/1.2.3
  STEP_TARGETS   build
)

If many external projects should have the same set of step targets, setting a directory property may be more convenient. The build step target could be created automatically by setting the EP_STEP_TARGETS directory property before creating the external projects with ExternalProject_Add():

set_property(DIRECTORY PROPERTY EP_STEP_TARGETS build)

Lastly, suppose that secretsauce provides a script called makedoc which can be used to generate its own documentation. Further suppose that the script expects the output directory to be provided as the only parameter and that it should be run from the secretsauce source directory. A custom step and a custom target to trigger the script can be defined like so:

ExternalProject_Add_Step(secretsauce docs
  COMMAND           <SOURCE_DIR>/makedoc <BINARY_DIR>
  WORKING_DIRECTORY <SOURCE_DIR>
  COMMENT           "Building secretsauce docs"
  ALWAYS            TRUE
  EXCLUDE_FROM_MAIN TRUE
)
ExternalProject_Add_StepTargets(secretsauce docs)

The custom step could then be triggered from the main build like so:

cmake --build . --target secretsauce-docs

Functions for generating a summary of enabled/disabled features.

These functions can be used to generate a summary of enabled and disabled packages and/or feature for a build tree such as:

-- The following OPTIONAL packages have been found:
LibXml2 (required version >= 2.4), XML processing lib, <http://xmlsoft.org>
   * Enables HTML-import in MyWordProcessor
   * Enables odt-export in MyWordProcessor
PNG, A PNG image library., <http://www.libpng.org/pub/png/>
   * Enables saving screenshots
-- The following OPTIONAL packages have not been found:
Lua51, The Lua scripting language., <https://www.lua.org>
   * Enables macros in MyWordProcessor
Foo, Foo provides cool stuff.

FeatureSummary_PKG_TYPES

The global property FeatureSummary_PKG_TYPES defines the type of packages used by FeatureSummary.

The order in this list is important, the first package type in the list is the least important, the last is the most important. the of a package can only be changed to higher types.

The default package types are , RUNTIME, OPTIONAL, RECOMMENDED and REQUIRED, and their importance is RUNTIME < OPTIONAL < RECOMMENDED < REQUIRED.

FeatureSummary_REQUIRED_PKG_TYPES

The global property FeatureSummary_REQUIRED_PKG_TYPES defines which package types are required.

If one or more package in this categories has not been found, CMake will abort when calling feature_summary() with the 'FATAL_ON_MISSING_REQUIRED_PACKAGES' option enabled.

The default value for this global property is REQUIRED.

FeatureSummary_DEFAULT_PKG_TYPE

The global property FeatureSummary_DEFAULT_PKG_TYPE defines which package type is the default one. When calling feature_summary(), if the user did not set the package type explicitly, the package will be assigned to this category.

This value must be one of the types defined in the FeatureSummary_PKG_TYPES global property unless the package type is set for all the packages.

The default value for this global property is OPTIONAL.

FeatureSummary_<TYPE>_DESCRIPTION

Added in version 3.9.

The global property FeatureSummary_<TYPE>_DESCRIPTION can be defined for each type to replace the type name with the specified string whenever the package type is used in an output string.

If not set, the string "<TYPE> packages" is used.

feature_summary

feature_summary( [FILENAME <file>]
                 [APPEND]
                 [VAR <variable_name>]
                 [INCLUDE_QUIET_PACKAGES]
                 [FATAL_ON_MISSING_REQUIRED_PACKAGES]
                 [DESCRIPTION "<description>" | DEFAULT_DESCRIPTION]
                 [QUIET_ON_EMPTY]
                 WHAT (ALL
                      | PACKAGES_FOUND | PACKAGES_NOT_FOUND
                      | <TYPE>_PACKAGES_FOUND | <TYPE>_PACKAGES_NOT_FOUND
                      | ENABLED_FEATURES | DISABLED_FEATURES)
               )

The feature_summary() macro can be used to print information about enabled or disabled packages or features of a project.  By default, only the names of the features/packages will be printed and their required version when one was specified.  Use set_package_properties() to add more useful information, like e.g.  a download URL for the respective package or their purpose in the project.

The WHAT option is the only mandatory option.  Here you specify what information will be printed:

ALL

print everything

ENABLED_FEATURES

the list of all features which are enabled

DISABLED_FEATURES

the list of all features which are disabled

PACKAGES_FOUND

the list of all packages which have been found

PACKAGES_NOT_FOUND

the list of all packages which have not been found

For each package type <TYPE> defined by the FeatureSummary_PKG_TYPES global property, the following information can also be used:

<TYPE>_PACKAGES_FOUND

only those packages which have been found which have the type <TYPE>

<TYPE>_PACKAGES_NOT_FOUND

only those packages which have not been found which have the type <TYPE>

Changed in version 3.1: With the exception of the ALL value, these values can be combined in order to customize the output. For example:

feature_summary(WHAT ENABLED_FEATURES DISABLED_FEATURES)

If a FILENAME is given, the information is printed into this file.  If APPEND is used, it is appended to this file, otherwise the file is overwritten if it already existed.  If the VAR option is used, the information is "printed" into the specified variable.  If FILENAME is not used, the information is printed to the terminal.  Using the DESCRIPTION option a description or headline can be set which will be printed above the actual content.  If only one type of package was requested, no title is printed, unless it is explicitly set using either DESCRIPTION to use a custom string, or DEFAULT_DESCRIPTION to use a default title for the requested type. If INCLUDE_QUIET_PACKAGES is given, packages which have been searched with find_package(... QUIET) will also be listed. By default they are skipped. If FATAL_ON_MISSING_REQUIRED_PACKAGES is given, CMake will abort if a package which is marked as one of the package types listed in the FeatureSummary_REQUIRED_PKG_TYPES global property has not been found. The default value for the FeatureSummary_REQUIRED_PKG_TYPES global property is REQUIRED.

Added in version 3.9: The DEFAULT_DESCRIPTION option.

The FeatureSummary_DEFAULT_PKG_TYPE global property can be modified to change the default package type assigned when not explicitly assigned by the user.

Added in version 3.8: If the QUIET_ON_EMPTY option is used, if only one type of package was requested, and no packages belonging to that category were found, then no output (including the DESCRIPTION) is printed or added to the VAR variable.

Example 1, append everything to a file:

include(FeatureSummary)
feature_summary(WHAT ALL
                FILENAME ${CMAKE_BINARY_DIR}/all.log APPEND)

Example 2, print the enabled features into the variable enabledFeaturesText, including QUIET packages:

include(FeatureSummary)
feature_summary(WHAT ENABLED_FEATURES
                INCLUDE_QUIET_PACKAGES
                DESCRIPTION "Enabled Features:"
                VAR enabledFeaturesText)
message(STATUS "${enabledFeaturesText}")

Example 3, change default package types and print only the categories that are not empty:

include(FeatureSummary)
set_property(GLOBAL APPEND PROPERTY FeatureSummary_PKG_TYPES BUILD)
find_package(FOO)
set_package_properties(FOO PROPERTIES TYPE BUILD)
feature_summary(WHAT BUILD_PACKAGES_FOUND
                Description "Build tools found:"
                QUIET_ON_EMPTY)
feature_summary(WHAT BUILD_PACKAGES_NOT_FOUND
                Description "Build tools not found:"
                QUIET_ON_EMPTY)

set_package_properties

set_package_properties(<name> PROPERTIES
                       [ URL <url> ]
                       [ DESCRIPTION <description> ]
                       [ TYPE (RUNTIME|OPTIONAL|RECOMMENDED|REQUIRED) ]
                       [ PURPOSE <purpose> ]
                      )

Use this macro to set up information about the named package, which can then be displayed via FEATURE_SUMMARY().  This can be done either directly in the Find-module or in the project which uses the module after the find_package() call.  The features for which information can be set are added automatically by the find_package() command.

URL <url>

This should be the homepage of the package, or something similar. Ideally this is set already directly in the Find-module.

DESCRIPTION <description>

A short description what that package is, at most one sentence. Ideally this is set already directly in the Find-module.

TYPE <type>

What type of dependency has the using project on that package. Default is OPTIONAL.  In this case it is a package which can be used by the project when available at buildtime, but it also work without. RECOMMENDED is similar to OPTIONAL, i.e.  the project will build if the package is not present, but the functionality of the resulting binaries will be severely limited.  If a REQUIRED package is not available at buildtime, the project may not even build.  This can be combined with the FATAL_ON_MISSING_REQUIRED_PACKAGES argument for feature_summary().  Last, a RUNTIME package is a package which is actually not used at all during the build, but which is required for actually running the resulting binaries.  So if such a package is missing, the project can still be built, but it may not work later on. If set_package_properties() is called multiple times for the same package with different TYPEs, the TYPE is only changed to higher TYPEs (RUNTIME < OPTIONAL < RECOMMENDED < REQUIRED), lower TYPEs are ignored.  The TYPE property is project-specific, so it cannot be set by the Find-module, but must be set in the project. Type accepted can be changed by setting the FeatureSummary_PKG_TYPES global property.

PURPOSE <purpose>

This describes which features this package enables in the project, i.e.  it tells the user what functionality he gets in the resulting binaries.  If set_package_properties() is called multiple times for a package, all PURPOSE properties are appended to a list of purposes of the package in the project.  As the TYPE property, also the PURPOSE property is project-specific, so it cannot be set by the Find-module, but must be set in the project.

Example for setting the info for a package:

find_package(LibXml2)
set_package_properties(LibXml2 PROPERTIES
                       DESCRIPTION "A XML processing library."
                       URL "http://xmlsoft.org/")
# or
set_package_properties(LibXml2 PROPERTIES
                       TYPE RECOMMENDED
                       PURPOSE "Enables HTML-import in MyWordProcessor")
# or
set_package_properties(LibXml2 PROPERTIES
                       TYPE OPTIONAL
                       PURPOSE "Enables odt-export in MyWordProcessor")

find_package(DBUS)
set_package_properties(DBUS PROPERTIES
  TYPE RUNTIME
  PURPOSE "Necessary to disable the screensaver during a presentation")

add_feature_info

add_feature_info(<name> <enabled> <description>)

Use this macro to add information about a feature with the given <name>. <enabled> contains whether this feature is enabled or not. It can be a variable or a list of conditions. <description> is a text describing the feature.  The information can be displayed using feature_summary() for ENABLED_FEATURES and DISABLED_FEATURES respectively.

Changed in version 3.8: <enabled> can be a list of conditions.

Example for setting the info for a feature:

option(WITH_FOO "Help for foo" ON)
add_feature_info(Foo WITH_FOO "The Foo feature provides very cool stuff.")

The following macros are provided for compatibility with previous CMake versions:

set_package_info

set_package_info(<name> <description> [ <url> [<purpose>] ])

Use this macro to set up information about the named package, which can then be displayed via feature_summary().  This can be done either directly in the Find-module or in the project which uses the module after the find_package() call.  The features for which information can be set are added automatically by the find_package() command.

set_feature_info

set_feature_info(<name> <description> [<url>])

Does the same as:

set_package_info(<name> <description> <url>)

print_enabled_features

print_enabled_features()

Does the same as

feature_summary(WHAT ENABLED_FEATURES DESCRIPTION "Enabled features:")

print_disabled_features

print_disabled_features()

Does the same as

feature_summary(WHAT DISABLED_FEATURES DESCRIPTION "Disabled features:")

Added in version 3.11.

NOTE:

This module enables populating content at configure time via any method supported by the ExternalProject module.  Whereas ExternalProject_Add() downloads at build time, the FetchContent module makes content available immediately, allowing the configure step to use the content in commands like add_subdirectory(), include() or file() operations.

Content population details should be defined separately from the command that performs the actual population.  This separation ensures that all the dependency details are defined before anything might try to use them to populate content.  This is particularly important in more complex project hierarchies where dependencies may be shared between multiple projects.

The following shows a typical example of declaring content details for some dependencies and then ensuring they are populated with a separate call:

FetchContent_Declare(
  googletest
  GIT_REPOSITORY https://github.com/google/googletest.git
  GIT_TAG        703bd9caab50b139428cea1aaff9974ebee5742e # release-1.10.0
)
FetchContent_Declare(
  myCompanyIcons
  URL      https://intranet.mycompany.com/assets/iconset_1.12.tar.gz
  URL_HASH MD5=5588a7b18261c20068beabfb4f530b87
)

FetchContent_MakeAvailable(googletest myCompanyIcons)

The FetchContent_MakeAvailable() command ensures the named dependencies have been populated, either by an earlier call, or by populating them itself.  When performing the population, it will also add them to the main build, if possible, so that the main build can use the populated projects' targets, etc.  See the command's documentation for how these steps are performed.

When using a hierarchical project arrangement, projects at higher levels in the hierarchy are able to override the declared details of content specified anywhere lower in the project hierarchy.  The first details to be declared for a given dependency take precedence, regardless of where in the project hierarchy that occurs.  Similarly, the first call that tries to populate a dependency "wins", with subsequent populations reusing the result of the first instead of repeating the population again. See the Examples which demonstrate this scenario.

The FetchContent module also supports defining and populating content in a single call, with no check for whether the content has been populated elsewhere already.  This should not be done in projects, but may be appropriate for populating content in CMake script mode. See FetchContent_Populate() for details.

FetchContent_Declare

FetchContent_Declare(
  <name>
  <contentOptions>...
  [EXCLUDE_FROM_ALL]
  [SYSTEM]
  [OVERRIDE_FIND_PACKAGE |
   FIND_PACKAGE_ARGS args...]
)

The FetchContent_Declare() function records the options that describe how to populate the specified content.  If such details have already been recorded earlier in this project (regardless of where in the project hierarchy), this and all later calls for the same content <name> are ignored.  This "first to record, wins" approach is what allows hierarchical projects to have parent projects override content details of child projects.

The content <name> can be any string without spaces, but good practice would be to use only letters, numbers, and underscores.  The name will be treated case-insensitively, and it should be obvious for the content it represents. It is often the name of the child project, or the value given to its top level project() command (if it is a CMake project). For well-known public projects, the name should generally be the official name of the project.  Choosing an unusual name makes it unlikely that other projects needing that same content will use the same name, leading to the content being populated multiple times.

The <contentOptions> can be any of the download, update, or patch options that the ExternalProject_Add() command understands.  The configure, build, install, and test steps are explicitly disabled, so options related to those steps will be ignored.  The SOURCE_SUBDIR option is an exception, see FetchContent_MakeAvailable() for details on how that affects behavior.

Changed in version 3.30: When policy CMP0168 is set to NEW, some output-related and directory-related options are ignored.  See the policy documentation for details.

In most cases, <contentOptions> will just be a couple of options defining the download method and method-specific details like a commit tag or archive hash.  For example:

FetchContent_Declare(
  googletest
  GIT_REPOSITORY https://github.com/google/googletest.git
  GIT_TAG        703bd9caab50b139428cea1aaff9974ebee5742e # release-1.10.0
)

FetchContent_Declare(
  myCompanyIcons
  URL      https://intranet.mycompany.com/assets/iconset_1.12.tar.gz
  URL_HASH MD5=5588a7b18261c20068beabfb4f530b87
)

FetchContent_Declare(
  myCompanyCertificates
  SVN_REPOSITORY svn+ssh://svn.mycompany.com/srv/svn/trunk/certs
  SVN_REVISION   -r12345
)

Where contents are being fetched from a remote location and you do not control that server, it is advisable to use a hash for GIT_TAG rather than a branch or tag name.  A commit hash is more secure and helps to confirm that the downloaded contents are what you expected.

Changed in version 3.14: Commands for the download, update, or patch steps can access the terminal. This may be needed for things like password prompts or real-time display of command progress.

Added in version 3.22: The CMAKE_TLS_VERIFY, CMAKE_TLS_CAINFO, CMAKE_NETRC, and CMAKE_NETRC_FILE variables now provide the defaults for their corresponding content options, just like they do for ExternalProject_Add(). Previously, these variables were ignored by the FetchContent module.

Added in version 3.24:

FIND_PACKAGE_ARGS

This option is for scenarios where the FetchContent_MakeAvailable() command may first try a call to find_package() to satisfy the dependency for <name>. By default, such a call would be simply find_package(<name>), but FIND_PACKAGE_ARGS can be used to provide additional arguments to be appended after the <name>.  FIND_PACKAGE_ARGS can also be given with nothing after it, which indicates that find_package() can still be called if FETCHCONTENT_TRY_FIND_PACKAGE_MODE is set to OPT_IN, or is not set.

It would not normally be appropriate to specify REQUIRED as one of the additional arguments after FIND_PACKAGE_ARGS.  Doing so would mean the find_package() call must succeed, so none of the other details specified in the FetchContent_Declare() call would get a chance to be used as a fall-back.

Everything after the FIND_PACKAGE_ARGS keyword is appended to the find_package() call, so all other <contentOptions> must come before the FIND_PACKAGE_ARGS keyword.  If the CMAKE_FIND_PACKAGE_TARGETS_GLOBAL variable is set to true at the time FetchContent_Declare() is called, a GLOBAL keyword will be appended to the find_package() arguments if it was not already specified.  It will also be appended if FIND_PACKAGE_ARGS was not given, but FETCHCONTENT_TRY_FIND_PACKAGE_MODE was set to ALWAYS.

OVERRIDE_FIND_PACKAGE cannot be used when FIND_PACKAGE_ARGS is given.

Dependency Providers discusses another way that FetchContent_MakeAvailable() calls can be redirected. FIND_PACKAGE_ARGS is intended for project control, whereas dependency providers allow users to override project behavior.

OVERRIDE_FIND_PACKAGE

When a FetchContent_Declare(<name> ...) call includes this option, subsequent calls to find_package(<name> ...) will ensure that FetchContent_MakeAvailable(<name>) has been called, then use the config package files in the CMAKE_FIND_PACKAGE_REDIRECTS_DIR directory (which are usually created by FetchContent_MakeAvailable()). This effectively makes FetchContent_MakeAvailable() override find_package() for the named dependency, allowing the former to satisfy the package requirements of the latter.  FIND_PACKAGE_ARGS cannot be used when OVERRIDE_FIND_PACKAGE is given.

If a dependency provider has been set and the project calls find_package() for the <name> dependency, OVERRIDE_FIND_PACKAGE will not prevent the provider from seeing that call.  Dependency providers always have the opportunity to intercept any direct call to find_package(), except if that call contains the BYPASS_PROVIDER option.

Added in version 3.25:

SYSTEM

If the SYSTEM argument is provided, the SYSTEM directory property of a subdirectory added by FetchContent_MakeAvailable() will be set to true.  This will affect non-imported targets created as part of that command. See the SYSTEM target property documentation for a more detailed discussion of the effects.

Added in version 3.28:

EXCLUDE_FROM_ALL

If the EXCLUDE_FROM_ALL argument is provided, then targets in the subdirectory added by FetchContent_MakeAvailable() will not be included in the ALL target by default, and may be excluded from IDE project files. See the documentation for the directory property EXCLUDE_FROM_ALL for a detailed discussion of the effects.

FetchContent_MakeAvailable

Added in version 3.14.

FetchContent_MakeAvailable(<name1> [<name2>...])

This command ensures that each of the named dependencies are made available to the project by the time it returns.  There must have been a call to FetchContent_Declare() for each dependency, and the first such call will control how that dependency will be made available, as described below.

If <lowercaseName>_SOURCE_DIR is not set:

• Added in version 3.24: If a dependency provider is set, call the provider's command with FETCHCONTENT_MAKEAVAILABLE_SERIAL as the first argument, followed by the arguments of the first call to FetchContent_Declare() for <name>.  If SOURCE_DIR or BINARY_DIR were not part of the original declared arguments, they will be added with their default values. If FETCHCONTENT_TRY_FIND_PACKAGE_MODE was set to NEVER when the details were declared, any FIND_PACKAGE_ARGS will be omitted.  The OVERRIDE_FIND_PACKAGE keyword is also always omitted. If the provider fulfilled the request, FetchContent_MakeAvailable() will consider that dependency handled, skip the remaining steps below, and move on to the next dependency in the list.
• Added in version 3.24: If permitted, find_package(<name> [<args>...]) will be called, where <args>... may be provided by the FIND_PACKAGE_ARGS option in FetchContent_Declare(). The value of the FETCHCONTENT_TRY_FIND_PACKAGE_MODE variable at the time FetchContent_Declare() was called determines whether FetchContent_MakeAvailable() can call find_package(). If the CMAKE_FIND_PACKAGE_TARGETS_GLOBAL variable is set to true when FetchContent_MakeAvailable() is called, it still affects any imported targets created when that in turn calls find_package(), even if that variable was false when the corresponding details were declared.

If the dependency was not satisfied by a provider or a find_package() call, FetchContent_MakeAvailable() then uses the following logic to make the dependency available:

• If the dependency has already been populated earlier in this run, set the <lowercaseName>_POPULATED, <lowercaseName>_SOURCE_DIR, and <lowercaseName>_BINARY_DIR variables in the same way as a call to FetchContent_GetProperties(), then skip the remaining steps below and move on to the next dependency in the list.
• Populate the dependency using the details recorded by an earlier call to FetchContent_Declare(). Halt with a fatal error if no such details have been recorded. FETCHCONTENT_SOURCE_DIR_<uppercaseName> can be used to override the declared details and use content provided at the specified location instead.
• Added in version 3.24: Ensure the CMAKE_FIND_PACKAGE_REDIRECTS_DIR directory contains a <lowercaseName>-config.cmake and a <lowercaseName>-config-version.cmake file (or equivalently, <name>Config.cmake and <name>ConfigVersion.cmake). The directory that the CMAKE_FIND_PACKAGE_REDIRECTS_DIR variable points to is cleared at the start of every CMake run. If no config file exists after populating the dependency in the previous step, a minimal one will be written which includes any <lowercaseName>-extra.cmake or <name>Extra.cmake file with the OPTIONAL flag (so the files can be missing and won't generate a warning).  Similarly, if no config version file exists, a very simple one will be written which sets PACKAGE_VERSION_COMPATIBLE and PACKAGE_VERSION_EXACT to true.  This ensures all future calls to find_package() for the dependency will use the redirected config file, regardless of any version requirements. CMake cannot automatically determine an arbitrary dependency's version, so it cannot set PACKAGE_VERSION. When a dependency is pulled in via add_subdirectory() in the next step, it may choose to overwrite the generated config version file in CMAKE_FIND_PACKAGE_REDIRECTS_DIR with one that also sets PACKAGE_VERSION. The dependency may also write a <lowercaseName>-extra.cmake or <name>Extra.cmake file to perform custom processing, or define any variables that their normal (installed) package config file would otherwise usually define (many projects don't do any custom processing or set any variables and therefore have no need to do this). If required, the main project can write these files instead if the dependency project doesn't do so.  This allows the main project to add missing details from older dependencies that haven't or can't be updated to support this functionality. See Integrating With find_package() for examples.

• If the top directory of the populated content contains a CMakeLists.txt file, call add_subdirectory() to add it to the main build. It is not an error for there to be no CMakeLists.txt file, which allows the command to be used for dependencies that make downloaded content available at a known location, but which do not need or support being added directly to the build.

  Added in version 3.18: The SOURCE_SUBDIR option can be given in the declared details to look somewhere below the top directory instead (i.e. the same way that SOURCE_SUBDIR is used by the ExternalProject_Add() command).  The path provided with SOURCE_SUBDIR must be relative, and it will be treated as relative to the top directory.  It can also point to a directory that does not contain a CMakeLists.txt file, or even to a directory that doesn't exist.  This can be used to avoid adding a project that contains a CMakeLists.txt file in its top directory.

  Added in version 3.25: If the SYSTEM keyword was included in the call to FetchContent_Declare(), the SYSTEM keyword will be added to the add_subdirectory() command.

  Added in version 3.28: If the EXCLUDE_FROM_ALL keyword was included in the call to FetchContent_Declare(), the EXCLUDE_FROM_ALL keyword will be added to the add_subdirectory() command.

  Added in version 3.29: CMAKE_EXPORT_FIND_PACKAGE_NAME is set to the dependency name before calling add_subdirectory().

Projects should aim to declare the details of all dependencies they might use before they call FetchContent_MakeAvailable() for any of them. This ensures that if any of the dependencies are also sub-dependencies of one or more of the others, the main project still controls the details that will be used (because it will declare them first before the dependencies get a chance to).  In the following code samples, assume that the uses_other dependency also uses FetchContent to add the other dependency internally:

# WRONG: Should declare all details first
FetchContent_Declare(uses_other ...)
FetchContent_MakeAvailable(uses_other)

FetchContent_Declare(other ...)    # Will be ignored, uses_other beat us to it
FetchContent_MakeAvailable(other)  # Would use details declared by uses_other

# CORRECT: All details declared first, so they will take priority
FetchContent_Declare(uses_other ...)
FetchContent_Declare(other ...)
FetchContent_MakeAvailable(uses_other other)

Note that CMAKE_VERIFY_INTERFACE_HEADER_SETS is explicitly set to false upon entry to FetchContent_MakeAvailable(), and is restored to its original value before the command returns.  Developers typically only want to verify header sets from the main project, not those from any dependencies.  This local manipulation of the CMAKE_VERIFY_INTERFACE_HEADER_SETS variable provides that intuitive behavior.  You can use variables like CMAKE_PROJECT_INCLUDE or CMAKE_PROJECT_<PROJECT-NAME>_INCLUDE to turn verification back on for all or some dependencies.  You can also set the VERIFY_INTERFACE_HEADER_SETS property of individual targets.

FetchContent_Populate

The FetchContent_Populate() command is a self-contained call which can be used to perform content population as an isolated operation. It is rarely the right command to use, projects should almost always use FetchContent_Declare() and FetchContent_MakeAvailable() instead. The main use case for FetchContent_Populate() is in CMake script mode as part of implementing some other higher level custom feature.

FetchContent_Populate(
  <name>
  [QUIET]
  [SUBBUILD_DIR <subBuildDir>]
  [SOURCE_DIR <srcDir>]
  [BINARY_DIR <binDir>]
  ...
)

At least one option must be specified after <name>, otherwise the call is interpreted differently (see below). The supported options for FetchContent_Populate() are the same as those for FetchContent_Declare(), with a few exceptions. The following do not relate to populating content with FetchContent_Populate() and therefore are not supported:

• EXCLUDE_FROM_ALL
• SYSTEM
• OVERRIDE_FIND_PACKAGE
• FIND_PACKAGE_ARGS

The few options shown in the signature above are either specific to FetchContent_Populate(), or their behavior is slightly modified from how ExternalProject_Add() treats them:

QUIET

The QUIET option can be given to hide the output associated with populating the specified content.  If the population fails, the output will be shown regardless of whether this option was given or not so that the cause of the failure can be diagnosed.  The FETCHCONTENT_QUIET variable has no effect on FetchContent_Populate() calls of this form where the content details are provided directly.

Changed in version 3.30: The QUIET option and FETCHCONTENT_QUIET variable have no effect when policy CMP0168 is set to NEW. The output is still quiet by default in that case, but verbosity is controlled by the message logging level (see CMAKE_MESSAGE_LOG_LEVEL and --log-level).

SUBBUILD_DIR

The SUBBUILD_DIR argument can be provided to change the location of the sub-build created to perform the population.  The default value is ${CMAKE_CURRENT_BINARY_DIR}/<lowercaseName>-subbuild, and it would be unusual to need to override this default.  If a relative path is specified, it will be interpreted as relative to CMAKE_CURRENT_BINARY_DIR. This option should not be confused with the SOURCE_SUBDIR option, which only affects the FetchContent_MakeAvailable() command.

Changed in version 3.30: SUBBUILD_DIR is ignored when policy CMP0168 is set to NEW, since there is no sub-build in that case.

SOURCE_DIR, BINARY_DIR

The SOURCE_DIR and BINARY_DIR arguments are supported by ExternalProject_Add(), but different default values are used by FetchContent_Populate().  SOURCE_DIR defaults to ${CMAKE_CURRENT_BINARY_DIR}/<lowercaseName>-src, and BINARY_DIR defaults to ${CMAKE_CURRENT_BINARY_DIR}/<lowercaseName>-build. If a relative path is specified, it will be interpreted as relative to CMAKE_CURRENT_BINARY_DIR.

In addition to the above explicit options, any other unrecognized options are passed through unmodified to ExternalProject_Add() to set up the download, patch, and update steps.  The following options are explicitly prohibited (they are disabled by the FetchContent_Populate() command):

• CONFIGURE_COMMAND
• BUILD_COMMAND
• INSTALL_COMMAND
• TEST_COMMAND

With this form, the FETCHCONTENT_FULLY_DISCONNECTED and FETCHCONTENT_UPDATES_DISCONNECTED variables and policy CMP0170 are ignored.

When this form of FetchContent_Populate() returns, the following variables will be set in the scope of the caller:

<lowercaseName>_SOURCE_DIR

The location where the populated content can be found upon return.

<lowercaseName>_BINARY_DIR

A directory originally intended for use as a corresponding build directory, but is unlikely to be relevant when using this form of the command.

If using FetchContent_Populate() within CMake script mode, be aware that the implementation sets up a sub-build which therefore requires a CMake generator and build tool to be available. If these cannot be found by default, then the CMAKE_GENERATOR and potentially the CMAKE_MAKE_PROGRAM variables will need to be set appropriately on the command line invoking the script.

Changed in version 3.30: If policy CMP0168 is set to NEW, no sub-build is used. Within CMake script mode, that allows FetchContent_Populate() to be called without any build tool or CMake generator.

Added in version 3.18: Added support for the DOWNLOAD_NO_EXTRACT option.

FetchContent_Populate(<name>)

FetchContent_GetProperties

When using saved content details, a call to FetchContent_MakeAvailable() or FetchContent_Populate() records information in global properties which can be queried at any time. This information may include the source and binary directories associated with the content, and also whether or not the content population has been processed during the current configure run.

FetchContent_GetProperties(
  <name>
  [SOURCE_DIR <srcDirVar>]
  [BINARY_DIR <binDirVar>]
  [POPULATED <doneVar>]
)

The SOURCE_DIR, BINARY_DIR, and POPULATED options can be used to specify which properties should be retrieved.  Each option accepts a value which is the name of the variable in which to store that property.  Most of the time though, only <name> is given, in which case the call will then set the same variables as a call to FetchContent_MakeAvailable(name) or FetchContent_Populate(name). Note that the SOURCE_DIR and BINARY_DIR values can be empty if the call is fulfilled by a dependency provider.

This command is rarely needed when using FetchContent_MakeAvailable().  It is more commonly used as part of implementing the deprecated pattern with FetchContent_Populate(), which ensures that the relevant variables will always be defined regardless of whether or not the population has been performed elsewhere in the project already:

# WARNING: This pattern is deprecated, don't use it!
#
# Check if population has already been performed
FetchContent_GetProperties(depname)
if(NOT depname_POPULATED)
  # Fetch the content using previously declared details
  FetchContent_Populate(depname)

  # Set custom variables, policies, etc.
  # ...

  # Bring the populated content into the build
  add_subdirectory(${depname_SOURCE_DIR} ${depname_BINARY_DIR})
endif()

FetchContent_SetPopulated

Added in version 3.24.

NOTE:

This command should only be called by dependency providers.  Calling it in any other context is unsupported and future CMake versions may halt with a fatal error in such cases.

FetchContent_SetPopulated(
  <name>
  [SOURCE_DIR <srcDir>]
  [BINARY_DIR <binDir>]
)

If a provider command fulfills a FETCHCONTENT_MAKEAVAILABLE_SERIAL request, it must call this function before returning.  The SOURCE_DIR and BINARY_DIR arguments can be used to specify the values that FetchContent_GetProperties() should return for its corresponding arguments.  Only provide SOURCE_DIR and BINARY_DIR if they have the same meaning as if they had been populated by the built-in FetchContent_MakeAvailable() implementation.

A number of cache variables can influence the behavior where details from a FetchContent_Declare() call are used to populate content.

NOTE:

All of these variables are intended for the developer to customize behavior. They should not normally be set by the project.

FETCHCONTENT_BASE_DIR

In most cases, the saved details do not specify any options relating to the directories to use for the internal sub-build, final source, and build areas. It is generally best to leave these decisions up to the FetchContent module to handle on the project's behalf.  The FETCHCONTENT_BASE_DIR cache variable controls the point under which all content population directories are collected, but in most cases, developers would not need to change this.  The default location is ${CMAKE_BINARY_DIR}/_deps, but if developers change this value, they should aim to keep the path short and just below the top level of the build tree to avoid running into path length problems on Windows.

FETCHCONTENT_QUIET

The logging output during population can be quite verbose, making the configure stage quite noisy.  This cache option (ON by default) hides all population output unless an error is encountered.  If experiencing problems with hung downloads, temporarily switching this option off may help diagnose which content population is causing the issue.

Changed in version 3.30: FETCHCONTENT_QUIET is ignored if policy CMP0168 is set to NEW.  The output is still quiet by default in that case, but verbosity is controlled by the message logging level (see CMAKE_MESSAGE_LOG_LEVEL and --log-level).

FETCHCONTENT_FULLY_DISCONNECTED

When this option is enabled, no attempt is made to download or update any content.  It is assumed that all content has already been populated in a previous run, or the source directories have been pointed at existing contents the developer has provided manually (using options described further below).  When the developer knows that no changes have been made to any content details, turning this option ON can speed up the configure stage.  It is OFF by default.

NOTE:

The FETCHCONTENT_FULLY_DISCONNECTED variable is not an appropriate way to prevent any network access on the first run in a build directory. Doing so can break projects, lead to misleading error messages, and hide subtle population failures.  This variable is specifically intended to only be turned on after the first time CMake has been run. If you want to prevent network access even on the first run, use a dependency provider and populate the dependency from local content instead.

Changed in version 3.30: The constraint that the source directory has already been populated when FETCHCONTENT_FULLY_DISCONNECTED is true is now enforced. See policy CMP0170.

FETCHCONTENT_UPDATES_DISCONNECTED

This is a less severe download/update control compared to FETCHCONTENT_FULLY_DISCONNECTED.  Instead of bypassing all download and update logic, FETCHCONTENT_UPDATES_DISCONNECTED only prevents the update step from making connections to remote servers when using the git or hg download methods.  Updates still occur if details about the update step change, but the update is attempted with only the information already available locally (so switching to a different tag or commit that is already fetched locally will succeed, but switching to an unknown commit hash will fail).  The download step is not affected, so if content has not been downloaded previously, it will still be downloaded when this option is enabled.  This can speed up the configure step, but not as much as FETCHCONTENT_FULLY_DISCONNECTED. FETCHCONTENT_UPDATES_DISCONNECTED is OFF by default.

FETCHCONTENT_TRY_FIND_PACKAGE_MODE

Added in version 3.24.

This variable modifies the details that FetchContent_Declare() records for a given dependency.  While it ultimately controls the behavior of FetchContent_MakeAvailable(), it is the variable's value when FetchContent_Declare() is called that gets used.  It makes no difference what the variable is set to when FetchContent_MakeAvailable() is called.  Since the variable should only be set by the user and not by projects directly, it will typically have the same value throughout anyway, so this distinction is not usually noticeable.

FETCHCONTENT_TRY_FIND_PACKAGE_MODE ultimately controls whether FetchContent_MakeAvailable() is allowed to call find_package() to satisfy a dependency.  The variable can be set to one of the following values:

OPT_IN

FetchContent_MakeAvailable() will only call find_package() if the FetchContent_Declare() call included a FIND_PACKAGE_ARGS keyword.  This is also the default behavior if FETCHCONTENT_TRY_FIND_PACKAGE_MODE is not set.

ALWAYS

find_package() can be called by FetchContent_MakeAvailable() regardless of whether the FetchContent_Declare() call included a FIND_PACKAGE_ARGS keyword or not.  If no FIND_PACKAGE_ARGS keyword was given, the behavior will be as though FIND_PACKAGE_ARGS had been provided, with no additional arguments after it.

NEVER

FetchContent_MakeAvailable() will not call find_package().  Any FIND_PACKAGE_ARGS given to the FetchContent_Declare() call will be ignored.

As a special case, if the FETCHCONTENT_SOURCE_DIR_<uppercaseName> variable has a non-empty value for a dependency, it is assumed that the user is overriding all other methods of making that dependency available. FETCHCONTENT_TRY_FIND_PACKAGE_MODE will have no effect on that dependency and FetchContent_MakeAvailable() will not try to call find_package() for it.

In addition to the above, the following variables are also defined for each content name:

FETCHCONTENT_SOURCE_DIR_<uppercaseName>

If this is set, no download or update steps are performed for the specified content and the <lowercaseName>_SOURCE_DIR variable returned to the caller is pointed at this location.  This gives developers a way to have a separate checkout of the content that they can modify freely without interference from the build.  The build simply uses that existing source, but it still defines <lowercaseName>_BINARY_DIR to point inside its own build area.  Developers are strongly encouraged to use this mechanism rather than editing the sources populated in the default location, as changes to sources in the default location can be lost when content population details are changed by the project.

FETCHCONTENT_UPDATES_DISCONNECTED_<uppercaseName>

This is the per-content equivalent of FETCHCONTENT_UPDATES_DISCONNECTED.  If the global option or this option is ON, then updates for the git and hg methods will not contact any remote for the named content.  They will only use information already available locally.  Disabling updates for individual content can be useful for content whose details rarely change, while still leaving other frequently changing content with updates enabled.

This first fairly straightforward example ensures that some popular testing frameworks are available to the main build:

include(FetchContent)
FetchContent_Declare(
  googletest
  GIT_REPOSITORY https://github.com/google/googletest.git
  GIT_TAG        703bd9caab50b139428cea1aaff9974ebee5742e # release-1.10.0
)
FetchContent_Declare(
  Catch2
  GIT_REPOSITORY https://github.com/catchorg/Catch2.git
  GIT_TAG        605a34765aa5d5ecbf476b4598a862ada971b0cc # v3.0.1
)

# After the following call, the CMake targets defined by googletest and
# Catch2 will be available to the rest of the build
FetchContent_MakeAvailable(googletest Catch2)

For the previous example, if the user wanted to try to find googletest and Catch2 via find_package() first before trying to download and build them from source, they could set the FETCHCONTENT_TRY_FIND_PACKAGE_MODE variable to ALWAYS. This would also affect any other calls to FetchContent_Declare() throughout the project, which might not be acceptable.  The behavior can be enabled for just these two dependencies instead by adding FIND_PACKAGE_ARGS to the declared details and leaving FETCHCONTENT_TRY_FIND_PACKAGE_MODE unset, or set to OPT_IN:

include(FetchContent)
FetchContent_Declare(
  googletest
  GIT_REPOSITORY https://github.com/google/googletest.git
  GIT_TAG        703bd9caab50b139428cea1aaff9974ebee5742e # release-1.10.0
  FIND_PACKAGE_ARGS NAMES GTest
)
FetchContent_Declare(
  Catch2
  GIT_REPOSITORY https://github.com/catchorg/Catch2.git
  GIT_TAG        605a34765aa5d5ecbf476b4598a862ada971b0cc # v3.0.1
  FIND_PACKAGE_ARGS
)

# This will try calling find_package() first for both dependencies
FetchContent_MakeAvailable(googletest Catch2)

For Catch2, no additional arguments to find_package() are needed, so no additional arguments are provided after the FIND_PACKAGE_ARGS keyword.  For googletest, its package is more commonly called GTest, so arguments are added to support it being found by that name.

If the user wanted to disable FetchContent_MakeAvailable() from calling find_package() for any dependency, even if it provided FIND_PACKAGE_ARGS in its declared details, they could set FETCHCONTENT_TRY_FIND_PACKAGE_MODE to NEVER.

If the project wanted to indicate that these two dependencies should be downloaded and built from source and that find_package() calls should be redirected to use the built dependencies, the OVERRIDE_FIND_PACKAGE option should be used when declaring the content details:

include(FetchContent)
FetchContent_Declare(
  googletest
  GIT_REPOSITORY https://github.com/google/googletest.git
  GIT_TAG        703bd9caab50b139428cea1aaff9974ebee5742e # release-1.10.0
  OVERRIDE_FIND_PACKAGE
)
FetchContent_Declare(
  Catch2
  GIT_REPOSITORY https://github.com/catchorg/Catch2.git
  GIT_TAG        605a34765aa5d5ecbf476b4598a862ada971b0cc # v3.0.1
  OVERRIDE_FIND_PACKAGE
)

# The following will automatically forward through to FetchContent_MakeAvailable()
find_package(googletest)
find_package(Catch2)

CMake provides a FindGTest module which defines some variables that older projects may use instead of linking to the imported targets.  To support those cases, we can provide an extra file.  In keeping with the "first to define, wins" philosophy of FetchContent, we only write out that file if something else hasn't already done so.

FetchContent_MakeAvailable(googletest)

if(NOT EXISTS ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/googletest-extra.cmake AND
   NOT EXISTS ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/googletestExtra.cmake)
  file(WRITE ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/googletest-extra.cmake
[=[
if("${GTEST_LIBRARIES}" STREQUAL "" AND TARGET GTest::gtest)
  set(GTEST_LIBRARIES GTest::gtest)
endif()
if("${GTEST_MAIN_LIBRARIES}" STREQUAL "" AND TARGET GTest::gtest_main)
  set(GTEST_MAIN_LIBRARIES GTest::gtest_main)
endif()
if("${GTEST_BOTH_LIBRARIES}" STREQUAL "")
  set(GTEST_BOTH_LIBRARIES ${GTEST_LIBRARIES} ${GTEST_MAIN_LIBRARIES})
endif()
]=])
endif()

Projects will also likely be using find_package(GTest) rather than find_package(googletest), but it is possible to make use of the CMAKE_FIND_PACKAGE_REDIRECTS_DIR area to pull in the latter as a dependency of the former.  This is likely to be sufficient to satisfy a typical find_package(GTest) call.

FetchContent_MakeAvailable(googletest)

if(NOT EXISTS ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/gtest-config.cmake AND
   NOT EXISTS ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/GTestConfig.cmake)
  file(WRITE ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/gtest-config.cmake
[=[
include(CMakeFindDependencyMacro)
find_dependency(googletest)
]=])
endif()

if(NOT EXISTS ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/gtest-config-version.cmake AND
   NOT EXISTS ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/GTestConfigVersion.cmake)
  file(WRITE ${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/gtest-config-version.cmake
[=[
include(${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/googletest-config-version.cmake OPTIONAL)
if(NOT PACKAGE_VERSION_COMPATIBLE)
  include(${CMAKE_FIND_PACKAGE_REDIRECTS_DIR}/googletestConfigVersion.cmake OPTIONAL)
endif()
]=])
endif()

If the sub-project's CMakeLists.txt file is not at the top level of its source tree, the SOURCE_SUBDIR option can be used to tell FetchContent where to find it.  The following example shows how to use that option, and it also sets a variable which is meaningful to the subproject before pulling it into the main build (set as an INTERNAL cache variable to avoid problems with policy CMP0077):

include(FetchContent)
FetchContent_Declare(
  protobuf
  GIT_REPOSITORY https://github.com/protocolbuffers/protobuf.git
  GIT_TAG        ae50d9b9902526efd6c7a1907d09739f959c6297 # v3.15.0
  SOURCE_SUBDIR  cmake
)
set(protobuf_BUILD_TESTS OFF CACHE INTERNAL "")
FetchContent_MakeAvailable(protobuf)

In more complex project hierarchies, the dependency relationships can be more complicated.  Consider a hierarchy where projA is the top level project and it depends directly on projects projB and projC.  Both projB and projC can be built standalone and they also both depend on another project projD.  projB additionally depends on projE.  This example assumes that all five projects are available on a company git server.  The CMakeLists.txt of each project might have sections like the following:

projA

include(FetchContent)
FetchContent_Declare(
  projB
  GIT_REPOSITORY git@mycompany.com:git/projB.git
  GIT_TAG        4a89dc7e24ff212a7b5167bef7ab079d
)
FetchContent_Declare(
  projC
  GIT_REPOSITORY git@mycompany.com:git/projC.git
  GIT_TAG        4ad4016bd1d8d5412d135cf8ceea1bb9
)
FetchContent_Declare(
  projD
  GIT_REPOSITORY git@mycompany.com:git/projD.git
  GIT_TAG        origin/integrationBranch
)
FetchContent_Declare(
  projE
  GIT_REPOSITORY git@mycompany.com:git/projE.git
  GIT_TAG        v2.3-rc1
)

# Order is important, see notes in the discussion further below
FetchContent_MakeAvailable(projD projB projC)

projB

include(FetchContent)
FetchContent_Declare(
  projD
  GIT_REPOSITORY git@mycompany.com:git/projD.git
  GIT_TAG        20b415f9034bbd2a2e8216e9a5c9e632
)
FetchContent_Declare(
  projE
  GIT_REPOSITORY git@mycompany.com:git/projE.git
  GIT_TAG        68e20f674a48be38d60e129f600faf7d
)

FetchContent_MakeAvailable(projD projE)

projC

include(FetchContent)
FetchContent_Declare(
  projD
  GIT_REPOSITORY git@mycompany.com:git/projD.git
  GIT_TAG        7d9a17ad2c962aa13e2fbb8043fb6b8a
)

FetchContent_MakeAvailable(projD)

A few key points should be noted in the above:

• projB and projC define different content details for projD, but projA also defines a set of content details for projD. Because projA will define them first, the details from projB and projC will not be used.  The override details defined by projA are not required to match either of those from projB or projC, but it is up to the higher level project to ensure that the details it does define still make sense for the child projects.
• In the projA call to FetchContent_MakeAvailable(), projD is listed ahead of projB and projC, so it will be populated before either projB or projC. It isn't required for projA to do this, doing so ensures that projA fully controls the environment in which projD is brought into the build (directory properties are particularly relevant).
• While projA defines content details for projE, it does not need to explicitly call FetchContent_MakeAvailable(projE) or FetchContent_Populate(projD) itself.  Instead, it leaves that to the child projB.  For higher level projects, it is often enough to just define the override content details and leave the actual population to the child projects.  This saves repeating the same thing at each level of the project hierarchy unnecessarily, but it should only be done if directory properties set by dependencies are not expected to influence the population of the shared dependency (projE in this case).

Projects don't always need to add the populated content to the build. Sometimes the project just wants to make the downloaded content available at a predictable location.  The next example ensures that a set of standard company toolchain files (and potentially even the toolchain binaries themselves) is available early enough to be used for that same build.

cmake_minimum_required(VERSION 3.14)

include(FetchContent)
FetchContent_Declare(
  mycom_toolchains
  URL  https://intranet.mycompany.com//toolchains_1.3.2.tar.gz
)
FetchContent_MakeAvailable(mycom_toolchains)

project(CrossCompileExample)

The project could be configured to use one of the downloaded toolchains like so:

cmake -DCMAKE_TOOLCHAIN_FILE=_deps/mycom_toolchains-src/toolchain_arm.cmake /path/to/src

When CMake processes the CMakeLists.txt file, it will download and unpack the tarball into _deps/mycompany_toolchains-src relative to the build directory.  The CMAKE_TOOLCHAIN_FILE variable is not used until the project() command is reached, at which point CMake looks for the named toolchain file relative to the build directory.  Because the tarball has already been downloaded and unpacked by then, the toolchain file will be in place, even the very first time that cmake is run in the build directory.

This last example demonstrates how one might download and unpack a firmware tarball using CMake's script mode. The call to FetchContent_Populate() specifies all the content details and the unpacked firmware will be placed in a firmware directory below the current working directory.

getFirmware.cmake

# NOTE: Intended to be run in script mode with cmake -P
include(FetchContent)
FetchContent_Populate(
  firmware
  URL        https://mycompany.com/assets/firmware-1.23-arm.tar.gz
  URL_HASH   MD5=68247684da89b608d466253762b0ff11
  SOURCE_DIR firmware
)

This module provides functions intended to be used in Find Modules implementing find_package(<PackageName>) calls.

find_package_handle_standard_args

This command handles the REQUIRED, QUIET and version-related arguments of find_package().  It also sets the <PackageName>_FOUND variable.  The package is considered found if all variables listed contain valid results, e.g. valid filepaths.

There are two signatures:

find_package_handle_standard_args(<PackageName>
  (DEFAULT_MSG|<custom-failure-message>)
  <required-var>...
  )

find_package_handle_standard_args(<PackageName>
  [FOUND_VAR <result-var>]
  [REQUIRED_VARS <required-var>...]
  [VERSION_VAR <version-var>]
  [HANDLE_VERSION_RANGE]
  [HANDLE_COMPONENTS]
  [CONFIG_MODE]
  [NAME_MISMATCHED]
  [REASON_FAILURE_MESSAGE <reason-failure-message>]
  [FAIL_MESSAGE <custom-failure-message>]
  )

The <PackageName>_FOUND variable will be set to TRUE if all the variables <required-var>... are valid and any optional constraints are satisfied, and FALSE otherwise.  A success or failure message may be displayed based on the results and on whether the REQUIRED and/or QUIET option was given to the find_package() call.

The options are:

(DEFAULT_MSG|<custom-failure-message>)

In the simple signature this specifies the failure message. Use DEFAULT_MSG to ask for a default message to be computed (recommended).  Not valid in the full signature.

FOUND_VAR <result-var>

Deprecated since version 3.3.

Specifies either <PackageName>_FOUND or <PACKAGENAME>_FOUND as the result variable.  This exists only for compatibility with older versions of CMake and is now ignored. Result variables of both names are always set for compatibility.

REQUIRED_VARS <required-var>...

Specify the variables which are required for this package. These may be named in the generated failure message asking the user to set the missing variable values.  Therefore these should typically be cache entries such as FOO_LIBRARY and not output variables like FOO_LIBRARIES.

Changed in version 3.18: If HANDLE_COMPONENTS is specified, this option can be omitted.

VERSION_VAR <version-var>

Specify the name of a variable that holds the version of the package that has been found.  This version will be checked against the (potentially) specified required version given to the find_package() call, including its EXACT option. The default messages include information about the required version and the version which has been actually found, both if the version is ok or not.

HANDLE_VERSION_RANGE

Added in version 3.19.

Enable handling of a version range, if one is specified. Without this option, a developer warning will be displayed if a version range is specified.

HANDLE_COMPONENTS

Enable handling of package components.  In this case, the command will report which components have been found and which are missing, and the <PackageName>_FOUND variable will be set to FALSE if any of the required components (i.e. not the ones listed after the OPTIONAL_COMPONENTS option of find_package()) are missing.

CONFIG_MODE

Specify that the calling find module is a wrapper around a call to find_package(<PackageName> NO_MODULE).  This implies a VERSION_VAR value of <PackageName>_VERSION.  The command will automatically check whether the package configuration file was found.

REASON_FAILURE_MESSAGE <reason-failure-message>

Added in version 3.16.

Specify a custom message of the reason for the failure which will be appended to the default generated message.

FAIL_MESSAGE <custom-failure-message>

Specify a custom failure message instead of using the default generated message.  Not recommended.

NAME_MISMATCHED

Added in version 3.17.

Indicate that the <PackageName> does not match ${CMAKE_FIND_PACKAGE_NAME}. This is usually a mistake and raises a warning, but it may be intentional for usage of the command for components of a larger package.

Example for the simple signature:

find_package_handle_standard_args(LibXml2 DEFAULT_MSG
  LIBXML2_LIBRARY LIBXML2_INCLUDE_DIR)

The LibXml2 package is considered to be found if both LIBXML2_LIBRARY and LIBXML2_INCLUDE_DIR are valid. Then also LibXml2_FOUND is set to TRUE.  If it is not found and REQUIRED was used, it fails with a message(FATAL_ERROR), independent whether QUIET was used or not.  If it is found, success will be reported, including the content of the first <required-var>.  On repeated CMake runs, the same message will not be printed again.

NOTE:

Example for the full signature:

find_package_handle_standard_args(LibArchive
  REQUIRED_VARS LibArchive_LIBRARY LibArchive_INCLUDE_DIR
  VERSION_VAR LibArchive_VERSION)

In this case, the LibArchive package is considered to be found if both LibArchive_LIBRARY and LibArchive_INCLUDE_DIR are valid. Also the version of LibArchive will be checked by using the version contained in LibArchive_VERSION.  Since no FAIL_MESSAGE is given, the default messages will be printed.

Another example for the full signature:

find_package(Automoc4 QUIET NO_MODULE HINTS /opt/automoc4)
find_package_handle_standard_args(Automoc4  CONFIG_MODE)

In this case, a FindAutmoc4.cmake module wraps a call to find_package(Automoc4 NO_MODULE) and adds an additional search directory for automoc4.  Then the call to find_package_handle_standard_args produces a proper success/failure message.

find_package_check_version

Added in version 3.19.

Helper function which can be used to check if a <version> is valid against version-related arguments of find_package().

find_package_check_version(<version> <result-var>
  [HANDLE_VERSION_RANGE]
  [RESULT_MESSAGE_VARIABLE <message-var>]
  )

The <result-var> will hold a boolean value giving the result of the check.

The options are:

HANDLE_VERSION_RANGE

Enable handling of a version range, if one is specified. Without this option, a developer warning will be displayed if a version range is specified.

RESULT_MESSAGE_VARIABLE <message-var>

Specify a variable to get back a message describing the result of the check.

Example for the usage:

find_package_check_version(1.2.3 result HANDLE_VERSION_RANGE
  RESULT_MESSAGE_VARIABLE reason)
if (result)
  message (STATUS "${reason}")
else()
  message (FATAL_ERROR "${reason}")
endif()

find_package_message(<name> "message for user" "find result details")

This function is intended to be used in FindXXX.cmake modules files. It will print a message once for each unique find result.  This is useful for telling the user where a package was found.  The first argument specifies the name (XXX) of the package.  The second argument specifies the message to display.  The third argument lists details about the find result so that if they change the message will be displayed again.  The macro also obeys the QUIET argument to the find_package command.

Example:

if(X11_FOUND)
  find_package_message(X11 "Found X11: ${X11_X11_LIB}"
    "[${X11_X11_LIB}][${X11_INCLUDE_DIR}]")
else()
 ...
endif()

Fortran/C Interface Detection

This module automatically detects the API by which C and Fortran languages interact.

Variables that indicate if the mangling is found:

FortranCInterface_GLOBAL_FOUND

Global subroutines and functions.

FortranCInterface_MODULE_FOUND

Module subroutines and functions (declared by "MODULE PROCEDURE").

This module also provides the following variables to specify the detected mangling, though a typical use case does not need to reference them and can use the Module Functions below.

FortranCInterface_GLOBAL_PREFIX

Prefix for a global symbol without an underscore.

FortranCInterface_GLOBAL_SUFFIX

Suffix for a global symbol without an underscore.

FortranCInterface_GLOBAL_CASE

The case for a global symbol without an underscore, either UPPER or LOWER.

FortranCInterface_GLOBAL__PREFIX

Prefix for a global symbol with an underscore.

FortranCInterface_GLOBAL__SUFFIX

Suffix for a global symbol with an underscore.

FortranCInterface_GLOBAL__CASE

The case for a global symbol with an underscore, either UPPER or LOWER.

FortranCInterface_MODULE_PREFIX

Prefix for a module symbol without an underscore.

FortranCInterface_MODULE_MIDDLE

Middle of a module symbol without an underscore that appears between the name of the module and the name of the symbol.

FortranCInterface_MODULE_SUFFIX

Suffix for a module symbol without an underscore.

FortranCInterface_MODULE_CASE

The case for a module symbol without an underscore, either UPPER or LOWER.

FortranCInterface_MODULE__PREFIX

Prefix for a module symbol with an underscore.

FortranCInterface_MODULE__MIDDLE

Middle of a module symbol with an underscore that appears between the name of the module and the name of the symbol.

FortranCInterface_MODULE__SUFFIX

Suffix for a module symbol with an underscore.

FortranCInterface_MODULE__CASE

The case for a module symbol with an underscore, either UPPER or LOWER.

FortranCInterface_HEADER

The FortranCInterface_HEADER function is provided to generate a C header file containing macros to mangle symbol names:

FortranCInterface_HEADER(<file>
                         [MACRO_NAMESPACE <macro-ns>]
                         [SYMBOL_NAMESPACE <ns>]
                         [SYMBOLS [<module>:]<function> ...])

It generates in <file> definitions of the following macros:

#define FortranCInterface_GLOBAL (name,NAME) ...
#define FortranCInterface_GLOBAL_(name,NAME) ...
#define FortranCInterface_MODULE (mod,name, MOD,NAME) ...
#define FortranCInterface_MODULE_(mod,name, MOD,NAME) ...

These macros mangle four categories of Fortran symbols, respectively:

• Global symbols without '_': call mysub()
• Global symbols with '_'   : call my_sub()
• Module symbols without '_': use mymod; call mysub()
• Module symbols with '_'   : use mymod; call my_sub()

If mangling for a category is not known, its macro is left undefined. All macros require raw names in both lower case and upper case.

The options are:

MACRO_NAMESPACE

Replace the default FortranCInterface_ prefix with a given namespace <macro-ns>.

SYMBOLS

List symbols to mangle automatically with C preprocessor definitions:

<function>          ==> #define <ns><function> ...
<module>:<function> ==> #define <ns><module>_<function> ...

If the mangling for some symbol is not known then no preprocessor definition is created, and a warning is displayed.

SYMBOL_NAMESPACE

Prefix all preprocessor definitions generated by the SYMBOLS option with a given namespace <ns>.

FortranCInterface_VERIFY

The FortranCInterface_VERIFY function is provided to verify that the Fortran and C/C++ compilers work together:

FortranCInterface_VERIFY([CXX] [QUIET])

It tests whether a simple test executable using Fortran and C (and C++ when the CXX option is given) compiles and links successfully.  The result is stored in the cache entry FortranCInterface_VERIFIED_C (or FortranCInterface_VERIFIED_CXX if CXX is given) as a boolean. If the check fails and QUIET is not given the function terminates with a fatal error message describing the problem.  The purpose of this check is to stop a build early for incompatible compiler combinations.  The test is built in the Release configuration.

include(FortranCInterface)
FortranCInterface_HEADER(FC.h MACRO_NAMESPACE "FC_")

This creates a "FC.h" header that defines mangling macros FC_GLOBAL(), FC_GLOBAL_(), FC_MODULE(), and FC_MODULE_().

include(FortranCInterface)
FortranCInterface_HEADER(FCMangle.h
                         MACRO_NAMESPACE "FC_"
                         SYMBOL_NAMESPACE "FC_"
                         SYMBOLS mysub mymod:my_sub)

This creates a "FCMangle.h" header that defines the same FC_*() mangling macros as the previous example plus preprocessor symbols FC_mysub and FC_mymod_my_sub.

FortranCInterface is aware of possible GLOBAL and MODULE manglings for many Fortran compilers, but it also provides an interface to specify new possible manglings.  Set the variables:

FortranCInterface_GLOBAL_SYMBOLS
FortranCInterface_MODULE_SYMBOLS

before including FortranCInterface to specify manglings of the symbols MySub, My_Sub, MyModule:MySub, and My_Module:My_Sub. For example, the code:

set(FortranCInterface_GLOBAL_SYMBOLS mysub_ my_sub__ MYSUB_)
  #                                  ^^^^^  ^^^^^^   ^^^^^
set(FortranCInterface_MODULE_SYMBOLS
    __mymodule_MOD_mysub __my_module_MOD_my_sub)
  #   ^^^^^^^^     ^^^^^   ^^^^^^^^^     ^^^^^^
include(FortranCInterface)

tells FortranCInterface to try given GLOBAL and MODULE manglings. (The carets point at raw symbol names for clarity in this example but are not needed.)

Function for generation of export macros for libraries

This module provides the function GENERATE_EXPORT_HEADER().

Added in version 3.12: Added support for C projects.  Previous versions supported C++ project only.

The GENERATE_EXPORT_HEADER function can be used to generate a file suitable for preprocessor inclusion which contains EXPORT macros to be used in library classes:

GENERATE_EXPORT_HEADER( LIBRARY_TARGET
          [BASE_NAME <base_name>]
          [EXPORT_MACRO_NAME <export_macro_name>]
          [EXPORT_FILE_NAME <export_file_name>]
          [DEPRECATED_MACRO_NAME <deprecated_macro_name>]
          [NO_EXPORT_MACRO_NAME <no_export_macro_name>]
          [INCLUDE_GUARD_NAME <include_guard_name>]
          [STATIC_DEFINE <static_define>]
          [NO_DEPRECATED_MACRO_NAME <no_deprecated_macro_name>]
          [DEFINE_NO_DEPRECATED]
          [PREFIX_NAME <prefix_name>]
          [CUSTOM_CONTENT_FROM_VARIABLE <variable>]
)

The target properties CXX_VISIBILITY_PRESET and VISIBILITY_INLINES_HIDDEN can be used to add the appropriate compile flags for targets.  See the documentation of those target properties, and the convenience variables CMAKE_CXX_VISIBILITY_PRESET and CMAKE_VISIBILITY_INLINES_HIDDEN.

By default GENERATE_EXPORT_HEADER() generates macro names in a file name determined by the name of the library.  This means that in the simplest case, users of GenerateExportHeader will be equivalent to:

set(CMAKE_CXX_VISIBILITY_PRESET hidden)
set(CMAKE_VISIBILITY_INLINES_HIDDEN 1)
add_library(somelib someclass.cpp)
generate_export_header(somelib)
install(TARGETS somelib DESTINATION ${LIBRARY_INSTALL_DIR})
install(FILES
 someclass.h
 ${PROJECT_BINARY_DIR}/somelib_export.h DESTINATION ${INCLUDE_INSTALL_DIR}
)

And in the ABI header files:

#include "somelib_export.h"
class SOMELIB_EXPORT SomeClass {
  ...
};

The CMake fragment will generate a file in the ${CMAKE_CURRENT_BINARY_DIR} called somelib_export.h containing the macros SOMELIB_EXPORT, SOMELIB_NO_EXPORT, SOMELIB_DEPRECATED, SOMELIB_DEPRECATED_EXPORT and SOMELIB_DEPRECATED_NO_EXPORT. They will be followed by content taken from the variable specified by the CUSTOM_CONTENT_FROM_VARIABLE option, if any. The resulting file should be installed with other headers in the library.

The BASE_NAME argument can be used to override the file name and the names used for the macros:

add_library(somelib someclass.cpp)
generate_export_header(somelib
  BASE_NAME other_name
)

Generates a file called other_name_export.h containing the macros OTHER_NAME_EXPORT, OTHER_NAME_NO_EXPORT and OTHER_NAME_DEPRECATED etc.

The BASE_NAME may be overridden by specifying other options in the function.  For example:

add_library(somelib someclass.cpp)
generate_export_header(somelib
  EXPORT_MACRO_NAME OTHER_NAME_EXPORT
)

creates the macro OTHER_NAME_EXPORT instead of SOMELIB_EXPORT, but other macros and the generated file name is as default:

add_library(somelib someclass.cpp)
generate_export_header(somelib
  DEPRECATED_MACRO_NAME KDE_DEPRECATED
)

creates the macro KDE_DEPRECATED instead of SOMELIB_DEPRECATED.

If LIBRARY_TARGET is a static library, macros are defined without values.

If the same sources are used to create both a shared and a static library, the uppercased symbol ${BASE_NAME}_STATIC_DEFINE should be used when building the static library:

add_library(shared_variant SHARED ${lib_SRCS})
add_library(static_variant ${lib_SRCS})
generate_export_header(shared_variant BASE_NAME libshared_and_static)
set_target_properties(static_variant PROPERTIES
  COMPILE_FLAGS -DLIBSHARED_AND_STATIC_STATIC_DEFINE)

This will cause the export macros to expand to nothing when building the static library.

If DEFINE_NO_DEPRECATED is specified, then a macro ${BASE_NAME}_NO_DEPRECATED will be defined This macro can be used to remove deprecated code from preprocessor output:

option(EXCLUDE_DEPRECATED "Exclude deprecated parts of the library" FALSE)
if (EXCLUDE_DEPRECATED)
  set(NO_BUILD_DEPRECATED DEFINE_NO_DEPRECATED)
endif()
generate_export_header(somelib ${NO_BUILD_DEPRECATED})

And then in somelib:

class SOMELIB_EXPORT SomeClass
{
public:
#ifndef SOMELIB_NO_DEPRECATED
  SOMELIB_DEPRECATED void oldMethod();
#endif
};

#ifndef SOMELIB_NO_DEPRECATED
void SomeClass::oldMethod() {  }
#endif

If PREFIX_NAME is specified, the argument will be used as a prefix to all generated macros.

For example:

generate_export_header(somelib PREFIX_NAME VTK_)

Generates the macros VTK_SOMELIB_EXPORT etc.

Added in version 3.1: Library target can be an OBJECT library.

Added in version 3.7: Added the CUSTOM_CONTENT_FROM_VARIABLE option.

Added in version 3.11: Added the INCLUDE_GUARD_NAME option.

ADD_COMPILER_EXPORT_FLAGS( [<output_variable>] )

Deprecated since version 3.0: Set the target properties CXX_VISIBILITY_PRESET and VISIBILITY_INLINES_HIDDEN instead.

The ADD_COMPILER_EXPORT_FLAGS function adds -fvisibility=hidden to CMAKE_CXX_FLAGS if supported, and is a no-op on Windows which does not need extra compiler flags for exporting support. You may optionally pass a single argument to ADD_COMPILER_EXPORT_FLAGS that will be populated with the CXX_FLAGS required to enable visibility support for the compiler/architecture in use.

Define GNU standard installation directories

Provides install directory variables as defined by the GNU Coding Standards.

Inclusion of this module defines the following variables:

CMAKE_INSTALL_<dir>

Destination for files of a given type.  This value may be passed to the DESTINATION options of  install() commands for the corresponding file type.  It should be a path relative to the installation prefix so that it can be converted to an absolute path in a relocatable way.

While absolute paths are allowed, they are not recommended as they do not work with the cmake --install command's --prefix option, or with the cpack installer generators. In particular, there is no need to make paths absolute by prepending CMAKE_INSTALL_PREFIX; this prefix is used by default if the DESTINATION is a relative path.

CMAKE_INSTALL_FULL_<dir>

The absolute path generated from the corresponding CMAKE_INSTALL_<dir> value.  If the value is not already an absolute path, an absolute path is constructed typically by prepending the value of the CMAKE_INSTALL_PREFIX variable.  However, there are some special cases as documented below.

These variables shouldn't be used in install() commands as they do not work with the cmake --install command's --prefix option, or with the cpack installer generators.

where <dir> is one of:

BINDIR

user executables (bin)

SBINDIR

system admin executables (sbin)

LIBEXECDIR

program executables (libexec)

SYSCONFDIR

read-only single-machine data (etc)

SHAREDSTATEDIR

modifiable architecture-independent data (com)

LOCALSTATEDIR

modifiable single-machine data (var)

RUNSTATEDIR

Added in version 3.9: run-time variable data (LOCALSTATEDIR/run)

LIBDIR

object code libraries (lib or lib64)

On Debian, this may be lib/<multiarch-tuple> when CMAKE_INSTALL_PREFIX is /usr.

INCLUDEDIR

C header files (include)

OLDINCLUDEDIR

C header files for non-gcc (/usr/include)

DATAROOTDIR

read-only architecture-independent data root (share)

DATADIR

read-only architecture-independent data (DATAROOTDIR)

INFODIR

info documentation (DATAROOTDIR/info)

LOCALEDIR

locale-dependent data (DATAROOTDIR/locale)

MANDIR

man documentation (DATAROOTDIR/man)

DOCDIR

documentation root (DATAROOTDIR/doc/PROJECT_NAME)

If the includer does not define a value the above-shown default will be used and the value will appear in the cache for editing by the user.

Added in version 3.4.

The following values of CMAKE_INSTALL_PREFIX are special:

/

For <dir> other than the SYSCONFDIR, LOCALSTATEDIR and RUNSTATEDIR, the value of CMAKE_INSTALL_<dir> is prefixed with usr/ if it is not user-specified as an absolute path. For example, the INCLUDEDIR value include becomes usr/include. This is required by the GNU Coding Standards, which state:

When building the complete GNU system, the prefix will be empty and /usr will be a symbolic link to /.

/usr

For <dir> equal to SYSCONFDIR, LOCALSTATEDIR or RUNSTATEDIR, the CMAKE_INSTALL_FULL_<dir> is computed by prepending just / to the value of CMAKE_INSTALL_<dir> if it is not user-specified as an absolute path. For example, the SYSCONFDIR value etc becomes /etc. This is required by the GNU Coding Standards.

/opt/...

For <dir> equal to SYSCONFDIR, LOCALSTATEDIR or RUNSTATEDIR, the CMAKE_INSTALL_FULL_<dir> is computed by appending the prefix to the value of CMAKE_INSTALL_<dir> if it is not user-specified as an absolute path. For example, the SYSCONFDIR value etc becomes /etc/opt/.... This is defined by the Filesystem Hierarchy Standard.

This behavior does not apply to paths under /opt/homebrew/....

GNUInstallDirs_get_absolute_install_dir

GNUInstallDirs_get_absolute_install_dir(absvar var dirname)

Added in version 3.7.

Set the given variable absvar to the absolute path contained within the variable var.  This is to allow the computation of an absolute path, accounting for all the special cases documented above.  While this macro is used to compute the various CMAKE_INSTALL_FULL_<dir> variables, it is exposed publicly to allow users who create additional path variables to also compute absolute paths where necessary, using the same logic.  dirname is the directory name to get, e.g. BINDIR.

Changed in version 3.20: Added the <dirname> parameter.  Previous versions of CMake passed this value through the variable ${dir}.

Added in version 3.9.

This module defines functions to help use the Google Test infrastructure.  Two mechanisms for adding tests are provided. gtest_add_tests() has been around for some time, originally via find_package(GTest). gtest_discover_tests() was introduced in CMake 3.10.

The (older) gtest_add_tests() scans source files to identify tests. This is usually effective, with some caveats, including in cross-compiling environments, and makes setting additional properties on tests more convenient. However, its handling of parameterized tests is less comprehensive, and it requires re-running CMake to detect changes to the list of tests.

The (newer) gtest_discover_tests() discovers tests by asking the compiled test executable to enumerate its tests.  This is more robust and provides better handling of parameterized tests, and does not require CMake to be re-run when tests change.  However, it may not work in a cross-compiling environment, and setting test properties is less convenient.

More details can be found in the documentation of the respective functions.

Both commands are intended to replace use of add_test() to register tests, and will create a separate CTest test for each Google Test test case. Note that this is in some cases less efficient, as common set-up and tear-down logic cannot be shared by multiple test cases executing in the same instance. However, it provides more fine-grained pass/fail information to CTest, which is usually considered as more beneficial.  By default, the CTest test name is the same as the Google Test name (i.e. suite.testcase); see also TEST_PREFIX and TEST_SUFFIX.

gtest_add_tests

Automatically add tests with CTest by scanning source code for Google Test macros:

gtest_add_tests(TARGET target
                [SOURCES src1...]
                [EXTRA_ARGS args...]
                [WORKING_DIRECTORY dir]
                [TEST_PREFIX prefix]
                [TEST_SUFFIX suffix]
                [SKIP_DEPENDENCY]
                [TEST_LIST outVar]
)

gtest_add_tests attempts to identify tests by scanning source files. Although this is generally effective, it uses only a basic regular expression match, which can be defeated by atypical test declarations, and is unable to fully "split" parameterized tests.  Additionally, it requires that CMake be re-run to discover any newly added, removed or renamed tests (by default, this means that CMake is re-run when any test source file is changed, but see SKIP_DEPENDENCY).  However, it has the advantage of declaring tests at CMake time, which somewhat simplifies setting additional properties on tests, and always works in a cross-compiling environment.

The options are:

TARGET target

Specifies the Google Test executable, which must be a known CMake executable target.  CMake will substitute the location of the built executable when running the test.

SOURCES src1...

When provided, only the listed files will be scanned for test cases.  If this option is not given, the SOURCES property of the specified target will be used to obtain the list of sources.

EXTRA_ARGS args...

Any extra arguments to pass on the command line to each test case.

Changed in version 3.31: Empty values in args... are preserved, see CMP0178.

WORKING_DIRECTORY dir

Specifies the directory in which to run the discovered test cases.  If this option is not provided, the current binary directory is used.

TEST_PREFIX prefix

Specifies a prefix to be prepended to the name of each discovered test case.  This can be useful when the same source files are being used in multiple calls to gtest_add_test() but with different EXTRA_ARGS.

TEST_SUFFIX suffix

Similar to TEST_PREFIX except the suffix is appended to the name of every discovered test case.  Both TEST_PREFIX and TEST_SUFFIX may be specified.

SKIP_DEPENDENCY

Normally, the function creates a dependency which will cause CMake to be re-run if any of the sources being scanned are changed.  This is to ensure that the list of discovered tests is updated.  If this behavior is not desired (as may be the case while actually writing the test cases), this option can be used to prevent the dependency from being added.

TEST_LIST outVar

The variable named by outVar will be populated in the calling scope with the list of discovered test cases.  This allows the caller to do things like manipulate test properties of the discovered tests.

Changed in version 3.31: Empty values in the TEST_LAUNCHER and CROSSCOMPILING_EMULATOR target properties are preserved, see policy CMP0178.

Usage example:

include(GoogleTest)
add_executable(FooTest FooUnitTest.cxx)
gtest_add_tests(TARGET      FooTest
                TEST_SUFFIX .noArgs
                TEST_LIST   noArgsTests
)
gtest_add_tests(TARGET      FooTest
                EXTRA_ARGS  --someArg someValue
                TEST_SUFFIX .withArgs
                TEST_LIST   withArgsTests
)
set_tests_properties(${noArgsTests}   PROPERTIES TIMEOUT 10)
set_tests_properties(${withArgsTests} PROPERTIES TIMEOUT 20)

For backward compatibility, the following form is also supported:

gtest_add_tests(exe args files...)

exe

The path to the test executable or the name of a CMake target.

args

A ;-list of extra arguments to be passed to executable.  The entire list must be passed as a single argument.  Enclose it in quotes, or pass "" for no arguments.

files...

A list of source files to search for tests and test fixtures. Alternatively, use AUTO to specify that exe is the name of a CMake executable target whose sources should be scanned.

include(GoogleTest)
set(FooTestArgs --foo 1 --bar 2)
add_executable(FooTest FooUnitTest.cxx)
gtest_add_tests(FooTest "${FooTestArgs}" AUTO)

gtest_discover_tests

Automatically add tests with CTest by querying the compiled test executable for available tests:

gtest_discover_tests(target
                     [EXTRA_ARGS args...]
                     [WORKING_DIRECTORY dir]
                     [TEST_PREFIX prefix]
                     [TEST_SUFFIX suffix]
                     [TEST_FILTER expr]
                     [NO_PRETTY_TYPES] [NO_PRETTY_VALUES]
                     [PROPERTIES name1 value1...]
                     [TEST_LIST var]
                     [DISCOVERY_TIMEOUT seconds]
                     [XML_OUTPUT_DIR dir]
                     [DISCOVERY_MODE <POST_BUILD|PRE_TEST>]
                     [DISCOVERY_EXTRA_ARGS args...]
)

Added in version 3.10.

gtest_discover_tests() sets up a post-build or pre-test command on the test executable that generates the list of tests by parsing the output from running the test executable with the --gtest_list_tests argument. Compared to the source parsing approach of gtest_add_tests(), this ensures that the full list of tests, including instantiations of parameterized tests, is obtained.  Since test discovery occurs at build or test time, it is not necessary to re-run CMake when the list of tests changes.  However, it requires that CROSSCOMPILING_EMULATOR is properly set in order to function in a cross-compiling environment.

Additionally, setting properties on tests is somewhat less convenient, since the tests are not available at CMake time.  Additional test properties may be assigned to the set of tests as a whole using the PROPERTIES option.  If more fine-grained test control is needed, custom content may be provided through an external CTest script using the TEST_INCLUDE_FILES directory property.  The set of discovered tests is made accessible to such a script via the <target>_TESTS variable (see the TEST_LIST option below for further discussion and limitations).

The options are:

target

Specifies the Google Test executable, which must be a known CMake executable target.  CMake will substitute the location of the built executable when running the test.

EXTRA_ARGS args...

Any extra arguments to pass on the command line to each test case.

Changed in version 3.31: Empty values in args... are preserved, see CMP0178.

WORKING_DIRECTORY dir

Specifies the directory in which to run the discovered test cases.  If this option is not provided, the current binary directory is used.

TEST_PREFIX prefix

Specifies a prefix to be prepended to the name of each discovered test case.  This can be useful when the same test executable is being used in multiple calls to gtest_discover_tests() but with different EXTRA_ARGS.

TEST_SUFFIX suffix

Similar to TEST_PREFIX except the suffix is appended to the name of every discovered test case.  Both TEST_PREFIX and TEST_SUFFIX may be specified.

TEST_FILTER expr

Added in version 3.22.

Filter expression to pass as a --gtest_filter argument during test discovery.  Note that the expression is a wildcard-based format that matches against the original test names as used by gtest.  For type or value-parameterized tests, these names may be different to the potentially pretty-printed test names that ctest uses.

NO_PRETTY_TYPES

By default, the type index of type-parameterized tests is replaced by the actual type name in the CTest test name.  If this behavior is undesirable (e.g. because the type names are unwieldy), this option will suppress this behavior.

NO_PRETTY_VALUES

By default, the value index of value-parameterized tests is replaced by the actual value in the CTest test name.  If this behavior is undesirable (e.g. because the value strings are unwieldy), this option will suppress this behavior.

PROPERTIES name1 value1...

Specifies additional properties to be set on all tests discovered by this invocation of gtest_discover_tests().

TEST_LIST var

Make the list of tests available in the variable var, rather than the default <target>_TESTS.  This can be useful when the same test executable is being used in multiple calls to gtest_discover_tests(). Note that this variable is only available in CTest.

Due to a limitation of CMake's parsing rules, any test with a square bracket in its name will be omitted from the list of tests stored in this variable.  Such tests will still be defined and executed by ctest as normal though.

DISCOVERY_TIMEOUT num

Added in version 3.10.3.

Specifies how long (in seconds) CMake will wait for the test to enumerate available tests.  If the test takes longer than this, discovery (and your build) will fail.  Most test executables will enumerate their tests very quickly, but under some exceptional circumstances, a test may require a longer timeout.  The default is 5.  See also the TIMEOUT option of execute_process().

NOTE:

In CMake versions 3.10.1 and 3.10.2, this option was called TIMEOUT. This clashed with the TIMEOUT test property, which is one of the common properties that would be set with the PROPERTIES keyword, usually leading to legal but unintended behavior.  The keyword was changed to DISCOVERY_TIMEOUT in CMake 3.10.3 to address this problem.  The ambiguous behavior of the TIMEOUT keyword in 3.10.1 and 3.10.2 has not been preserved.

XML_OUTPUT_DIR dir

Added in version 3.18.

If specified, the parameter is passed along with --gtest_output=xml: to test executable. The actual file name is the same as the test target, including prefix and suffix. This should be used instead of EXTRA_ARGS --gtest_output=xml to avoid race conditions writing the XML result output when using parallel test execution.

DISCOVERY_MODE

Added in version 3.18.

Provides greater control over when gtest_discover_tests() performs test discovery. By default, POST_BUILD sets up a post-build command to perform test discovery at build time. In certain scenarios, like cross-compiling, this POST_BUILD behavior is not desirable. By contrast, PRE_TEST delays test discovery until just prior to test execution. This way test discovery occurs in the target environment where the test has a better chance at finding appropriate runtime dependencies.

DISCOVERY_MODE defaults to the value of the CMAKE_GTEST_DISCOVER_TESTS_DISCOVERY_MODE variable if it is not passed when calling gtest_discover_tests(). This provides a mechanism for globally selecting a preferred test discovery behavior without having to modify each call site.

DISCOVERY_EXTRA_ARGS args...

Added in version 3.31.

Any extra arguments to pass on the command line for the discovery command.

Added in version 3.29: The TEST_LAUNCHER target property is honored during test discovery and test execution.

Changed in version 3.31: Empty values in the TEST_LAUNCHER and CROSSCOMPILING_EMULATOR target properties are preserved, see policy CMP0178.

Include this module to search for compiler-provided system runtime libraries and add install rules for them.  Some optional variables may be set prior to including the module to adjust behavior:

CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS

Specify additional runtime libraries that may not be detected. After inclusion any detected libraries will be appended to this.

CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS_SKIP

Set to TRUE to skip calling the install(PROGRAMS) command to allow the includer to specify its own install rule, using the value of CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS to get the list of libraries.

CMAKE_INSTALL_DEBUG_LIBRARIES

Set to TRUE to install the debug runtime libraries when available with MSVC tools.

CMAKE_INSTALL_DEBUG_LIBRARIES_ONLY

Set to TRUE to install only the debug runtime libraries with MSVC tools even if the release runtime libraries are also available.

CMAKE_INSTALL_UCRT_LIBRARIES

Added in version 3.6.

Set to TRUE to install the Windows Universal CRT libraries for app-local deployment (e.g. to Windows XP).  This is meaningful only with MSVC from Visual Studio 2015 or higher.

Added in version 3.9: One may set a CMAKE_WINDOWS_KITS_10_DIR environment variable to an absolute path to tell CMake to look for Windows 10 SDKs in a custom location.  The specified directory is expected to contain Redist/ucrt/DLLs/* directories.

CMAKE_INSTALL_MFC_LIBRARIES

Set to TRUE to install the MSVC MFC runtime libraries.

CMAKE_INSTALL_OPENMP_LIBRARIES

Set to TRUE to install the MSVC OpenMP runtime libraries

CMAKE_INSTALL_SYSTEM_RUNTIME_DESTINATION

Specify the install(PROGRAMS) command DESTINATION option.  If not specified, the default is bin on Windows and lib elsewhere.

CMAKE_INSTALL_SYSTEM_RUNTIME_LIBS_NO_WARNINGS

Set to TRUE to disable warnings about required library files that do not exist.  (For example, Visual Studio Express editions may not provide the redistributable files.)

CMAKE_INSTALL_SYSTEM_RUNTIME_COMPONENT

Added in version 3.3.

Specify the install(PROGRAMS) command COMPONENT option.  If not specified, no such option will be used.

Added in version 3.10: Support for installing Intel compiler runtimes.

ProcessorCount(var)

Determine the number of processors/cores and save value in ${var}

Sets the variable named ${var} to the number of physical cores available on the machine if the information can be determined. Otherwise it is set to 0.  Currently this functionality is implemented for AIX, cygwin, FreeBSD, HPUX, Linux, macOS, QNX, Sun and Windows.

Changed in version 3.15: On Linux, returns the container CPU count instead of the host CPU count.

This function is guaranteed to return a positive integer (>=1) if it succeeds.  It returns 0 if there's a problem determining the processor count.

More generally accurate physical CPU count can be obtained via cmake_host_system_information():

cmake_host_system_information(RESULT N
                              QUERY NUMBER_OF_PHYSICAL_CORES)

Example use, in a ctest -S dashboard script:

include(ProcessorCount)
ProcessorCount(N)
if(NOT N EQUAL 0)
  set(CTEST_BUILD_FLAGS -j${N})
  set(ctest_test_args ${ctest_test_args} PARALLEL_LEVEL ${N})
endif()

This function is intended to offer an approximation of the value of the number of compute cores available on the current machine, such that you may use that value for parallel building and parallel testing.  It is meant to help utilize as much of the machine as seems reasonable.  Of course, knowledge of what else might be running on the machine simultaneously should be used when deciding whether to request a machine's full capacity all for yourself.

select_library_configurations(basename)

This macro takes a library base name as an argument, and will choose good values for the variables

basename_LIBRARY
basename_LIBRARIES
basename_LIBRARY_DEBUG
basename_LIBRARY_RELEASE

depending on what has been found and set.

If only basename_LIBRARY_RELEASE is defined, basename_LIBRARY will be set to the release value, and basename_LIBRARY_DEBUG will be set to basename_LIBRARY_DEBUG-NOTFOUND.  If only basename_LIBRARY_DEBUG is defined, then basename_LIBRARY will take the debug value, and basename_LIBRARY_RELEASE will be set to basename_LIBRARY_RELEASE-NOTFOUND.

If the generator supports configuration types, then basename_LIBRARY and basename_LIBRARIES will be set with debug and optimized flags specifying the library to be used for the given configuration.  If no build type has been set or the generator in use does not support configuration types, then basename_LIBRARY and basename_LIBRARIES will take only the release value, or the debug value if the release one is not set.

This script launches a GUI test using Squish.  You should not call the script directly; instead, you should access it via the SQUISH_ADD_TEST macro that is defined in FindSquish.cmake.

This script starts the Squish server, launches the test on the client, and finally stops the squish server.  If any of these steps fail (including if the tests do not pass) then a fatal error is raised.

Check for ANSI for scope support

Check if the compiler restricts the scope of variables declared in a for-init-statement to the loop body.

CMAKE_NO_ANSI_FOR_SCOPE - holds result

Test for compiler support of ANSI stream headers iostream, etc.

check if the compiler supports the standard ANSI iostream header (without the .h)

CMAKE_NO_ANSI_STREAM_HEADERS - defined by the results

Test for compiler support of ANSI sstream header

check if the compiler supports the standard ANSI sstream header

CMAKE_NO_ANSI_STRING_STREAM - defined by the results

Test for std:: namespace support

check if the compiler supports std:: on stl classes

CMAKE_NO_STD_NAMESPACE - defined by the results

This module defines variables and macros required to build eCos application.

This file contains the following macros:

ECOS_ADD_INCLUDE_DIRECTORIES()

add the eCos include dirs

ECOS_ADD_EXECUTABLE(name source1 ... sourceN )

create an eCos executable

ECOS_ADJUST_DIRECTORY(VAR source1 ... sourceN )

adjusts the path of the source files and puts the result into VAR

Macros for selecting the toolchain:

ECOS_USE_ARM_ELF_TOOLS()

enable the ARM ELF toolchain for the directory where it is called

ECOS_USE_I386_ELF_TOOLS()

enable the i386 ELF toolchain for the directory where it is called

ECOS_USE_PPC_EABI_TOOLS()

enable the PowerPC toolchain for the directory where it is called

It contains the following variables:

ECOS_DEFINITIONS

ECOSCONFIG_EXECUTABLE

ECOS_CONFIG_FILE

defaults to ecos.ecc, if your eCos configuration file has a different name, adjust this variable for internal use only:

ECOS_ADD_TARGET_LIB

This file provides support for Java.  It is assumed that FindJava has already been loaded.  See FindJava for information on how to load Java into your CMake project.

Creating and Installing JARS
  add_jar (<target_name> [SOURCES] <source1> [<source2>...] ...)
  install_jar (<target_name> DESTINATION <destination> [COMPONENT <component>])
  install_jni_symlink (<target_name> DESTINATION <destination> [COMPONENT <component>])

Header Generation
  create_javah ((TARGET <target> | GENERATED_FILES <VAR>) CLASSES <class>... ...)

Exporting JAR Targets
  install_jar_exports (TARGETS <jars>... FILE <filename> DESTINATION <destination> ...)
  export_jars (TARGETS <jars>... [NAMESPACE <namespace>] FILE <filename>)

Finding JARs
  find_jar (<VAR> NAMES <name1> [<name2>...] [PATHS <path1> [<path2>... ENV <var>]] ...)

Creating Java Documentation
  create_javadoc (<VAR> (PACKAGES <pkg1> [<pkg2>...] | FILES <file1> [<file2>...]) ...)

add_jar

Creates a jar file containing java objects and, optionally, resources:

add_jar(<target_name>
        [SOURCES] <source1> [<source2>...] [<resource1>...]
        [RESOURCES NAMESPACE <ns1> <resource1>... [NAMESPACE <nsX> <resourceX>...]... ]
        [INCLUDE_JARS <jar1> [<jar2>...]]
        [ENTRY_POINT <entry>]
        [VERSION <version>]
        [MANIFEST <manifest>]
        [OUTPUT_NAME <name>]
        [OUTPUT_DIR <dir>]
        [GENERATE_NATIVE_HEADERS <target>
                                 [DESTINATION (<dir>|INSTALL <dir> [BUILD <dir>])]]
        )

This command creates a <target_name>.jar.  It compiles the given <source> files and adds the given <resource> files to the jar file.  Source files can be java files or listing files (prefixed by @).  If only resource files are given then just a jar file is created.

SOURCES

Compiles the specified source files and adds the result in the jar file.

Added in version 3.4: Support for response files, prefixed by @.

RESOURCES

Added in version 3.21.

Adds the named <resource> files to the jar by stripping the source file path and placing the file beneath <ns> within the jar.

For example:

RESOURCES NAMESPACE "/com/my/namespace" "a/path/to/resource.txt"

results in a resource accessible via /com/my/namespace/resource.txt within the jar.

Resources may be added without adjusting the namespace by adding them to the list of SOURCES (original behavior), in this case, resource paths must be relative to CMAKE_CURRENT_SOURCE_DIR.  Adding resources without using the RESOURCES parameter in out of source builds will almost certainly result in confusion.

NOTE:

Adding resources via the SOURCES parameter relies upon a hard-coded list of file extensions which are tested to determine whether they compile (e.g. File.java). SOURCES files which match the extensions are compiled. Files which do not match are treated as resources. To include uncompiled resources matching those file extensions use the RESOURCES parameter.

INCLUDE_JARS

The list of jars are added to the classpath when compiling the java sources and also to the dependencies of the target. INCLUDE_JARS also accepts other target names created by add_jar(). For backwards compatibility, jar files listed as sources are ignored (as they have been since the first version of this module).

ENTRY_POINT

Defines an entry point in the jar file.

VERSION

Adds a version to the target output name.

The following example will create a jar file with the name shibboleet-1.2.0.jar and will create a symlink shibboleet.jar pointing to the jar with the version information.

add_jar(shibboleet shibbotleet.java VERSION 1.2.0)

MANIFEST

Defines a custom manifest for the jar.

OUTPUT_NAME

Specify a different output name for the target.

OUTPUT_DIR

Sets the directory where the jar file will be generated. If not specified, CMAKE_CURRENT_BINARY_DIR is used as the output directory.

GENERATE_NATIVE_HEADERS

Added in version 3.11.

Generates native header files for methods declared as native. These files provide the connective glue that allow your Java and C code to interact. An INTERFACE target will be created for an easy usage of generated files. Sub-option DESTINATION can be used to specify the output directory for generated header files.

This option requires, at least, version 1.8 of the JDK.

For an optimum usage of this option, it is recommended to include module JNI before any call to add_jar(). The produced target for native headers can then be used to compile C/C++ sources with the target_link_libraries() command.

find_package(JNI)
add_jar(foo foo.java GENERATE_NATIVE_HEADERS foo-native)
add_library(bar bar.cpp)
target_link_libraries(bar PRIVATE foo-native)

Added in version 3.20: DESTINATION sub-option now supports the possibility to specify different output directories for BUILD and INSTALL steps. If BUILD directory is not specified, a default directory will be used.

To export the interface target generated by GENERATE_NATIVE_HEADERS option, sub-option INSTALL of DESTINATION is required:

add_jar(foo foo.java GENERATE_NATIVE_HEADERS foo-native
                     DESTINATION INSTALL include)
install(TARGETS foo-native EXPORT native)
install(DIRECTORY "$<TARGET_PROPERTY:foo-native,NATIVE_HEADERS_DIRECTORY>/"
        DESTINATION include)
install(EXPORT native DESTINATION /to/export NAMESPACE foo)

Some variables can be set to customize the behavior of add_jar() as well as the java compiler:

CMAKE_JAVA_COMPILE_FLAGS

Specify additional flags to java compiler.

CMAKE_JAVA_INCLUDE_PATH

Specify additional paths to the class path.

CMAKE_JNI_TARGET

If the target is a JNI library, sets this boolean variable to TRUE to enable creation of a JNI symbolic link (see also install_jni_symlink()).

CMAKE_JAR_CLASSES_PREFIX

If multiple jars should be produced from the same java source filetree, to prevent the accumulation of duplicate class files in subsequent jars, set/reset CMAKE_JAR_CLASSES_PREFIX prior to calling the add_jar():

set(CMAKE_JAR_CLASSES_PREFIX com/redhat/foo)
add_jar(foo foo.java)

set(CMAKE_JAR_CLASSES_PREFIX com/redhat/bar)
add_jar(bar bar.java)

The add_jar() function sets the following target properties on <target_name>:

INSTALL_FILES

The files which should be installed.  This is used by install_jar().

JNI_SYMLINK

The JNI symlink which should be installed.  This is used by install_jni_symlink().

JAR_FILE

The location of the jar file so that you can include it.

CLASSDIR

The directory where the class files can be found.  For example to use them with javah.

NATIVE_HEADERS_DIRECTORY

Added in version 3.20.

The directory where native headers are generated. Defined when option GENERATE_NATIVE_HEADERS is specified.

install_jar

This command installs the jar file to the given destination:

install_jar(<target_name> <destination>)
install_jar(<target_name> DESTINATION <destination> [COMPONENT <component>])

This command installs the <target_name> file to the given <destination>.  It should be called in the same scope as add_jar() or it will fail.

Added in version 3.4: The second signature with DESTINATION and COMPONENT options.

DESTINATION

Specify the directory on disk to which a file will be installed.

COMPONENT

Specify an installation component name with which the install rule is associated, such as "runtime" or "development".

The install_jar() command sets the following target properties on <target_name>:

INSTALL_DESTINATION

Holds the <destination> as described above, and is used by install_jar_exports().

install_jni_symlink

Installs JNI symlinks for target generated by add_jar():

install_jni_symlink(<target_name> <destination>)
install_jni_symlink(<target_name> DESTINATION <destination> [COMPONENT <component>])

This command installs the <target_name> JNI symlinks to the given <destination>.  It should be called in the same scope as add_jar() or it will fail.

Added in version 3.4: The second signature with DESTINATION and COMPONENT options.

DESTINATION

Specify the directory on disk to which a file will be installed.

COMPONENT

Specify an installation component name with which the install rule is associated, such as "runtime" or "development".

Utilize the following commands to create a JNI symbolic link:

set(CMAKE_JNI_TARGET TRUE)
add_jar(shibboleet shibbotleet.java VERSION 1.2.0)
install_jar(shibboleet ${LIB_INSTALL_DIR}/shibboleet)
install_jni_symlink(shibboleet ${JAVA_LIB_INSTALL_DIR})

create_javah

Added in version 3.4.

Generates C header files for java classes:

create_javah(TARGET <target> | GENERATED_FILES <VAR>
             CLASSES <class>...
             [CLASSPATH <classpath>...]
             [DEPENDS <depend>...]
             [OUTPUT_NAME <path>|OUTPUT_DIR <path>]
             )

Deprecated since version 3.11: This command will no longer be supported starting with version 10 of the JDK due to the suppression of javah tool. The add_jar(GENERATE_NATIVE_HEADERS) command should be used instead.

Create C header files from java classes. These files provide the connective glue that allow your Java and C code to interact.

There are two main signatures for create_javah().  The first signature returns generated files through variable specified by the GENERATED_FILES option.  For example:

create_javah(GENERATED_FILES files_headers
  CLASSES org.cmake.HelloWorld
  CLASSPATH hello.jar
)

The second signature for create_javah() creates a target which encapsulates header files generation. E.g.

create_javah(TARGET target_headers
  CLASSES org.cmake.HelloWorld
  CLASSPATH hello.jar
)

Both signatures share same options.

CLASSES

Specifies Java classes used to generate headers.

CLASSPATH

Specifies various paths to look up classes. Here .class files, jar files or targets created by command add_jar can be used.

DEPENDS

Targets on which the javah target depends.

OUTPUT_NAME

Concatenates the resulting header files for all the classes listed by option CLASSES into <path>.  Same behavior as option -o of javah tool.

OUTPUT_DIR

Sets the directory where the header files will be generated.  Same behavior as option -d of javah tool.  If not specified, CMAKE_CURRENT_BINARY_DIR is used as the output directory.

install_jar_exports

Added in version 3.7.

Installs a target export file:

install_jar_exports(TARGETS <jars>...
                    [NAMESPACE <namespace>]
                    FILE <filename>
                    DESTINATION <destination> [COMPONENT <component>])

This command installs a target export file <filename> for the named jar targets to the given <destination> directory.  Its function is similar to that of install(EXPORT).

TARGETS

List of targets created by add_jar() command.

NAMESPACE

Added in version 3.9.

The <namespace> value will be prepend to the target names as they are written to the import file.

FILE

Specify name of the export file.

DESTINATION

Specify the directory on disk to which a file will be installed.

COMPONENT

Specify an installation component name with which the install rule is associated, such as "runtime" or "development".

export_jars

Added in version 3.7.

Writes a target export file:

export_jars(TARGETS <jars>...
            [NAMESPACE <namespace>]
            FILE <filename>)

This command writes a target export file <filename> for the named <jars> targets.  Its function is similar to that of export().

TARGETS

List of targets created by add_jar() command.

NAMESPACE

Added in version 3.9.

The <namespace> value will be prepend to the target names as they are written to the import file.

FILE

Specify name of the export file.