: Some developers prefer to simply download a header-only library and manually create an INTERFACE target rather than letting the library's own CMakeLists.txt (which might include tests or benchmarks) clutter their project. Strategic Considerations FetchContent — CMake 4.3.2 Documentation
: If you are downloading a library that uses a different build system (like Make or Autotools), FetchContent_MakeAvailable would fail or do nothing useful.
Understanding why FetchContent_MakeAvailable is often the "wrong" tool for download-only tasks: cmake fetchcontent download only
: When your project needs large datasets, images, or configuration files that don't need to be "compiled".
: This command performs the raw download and extraction. It explicitly disables the configure, build, and install steps of the underlying ExternalProject engine, leaving you with just the source files. Common Use Cases for Download-Only : Some developers prefer to simply download a
The standard way to perform a "download-only" operation involves two steps: declaring the dependency and then populating it manually without adding its targets to your project.
include(FetchContent) # 1. Declare the dependency FetchContent_Declare( myData URL https://example.com ) # 2. Populate it manually (Download only) FetchContent_GetProperties(myData) if(NOT mydata_POPULATED) FetchContent_Populate(myData) # The content is now in ${mydata_SOURCE_DIR} # No add_subdirectory() is called automatically endif() Use code with caution. Key Differences in Commands : This command performs the raw download and extraction
: This is the modern, recommended way for dependencies you intend to build. It checks if the content is already populated, and if not, it downloads it and immediately calls add_subdirectory() . If your dependency has a CMakeLists.txt , it will be integrated into your project's build tree.
To download content without automatically adding it to your build system, use instead of the standard FetchContent_MakeAvailable command. While FetchContent_MakeAvailable is the "automatic" path that both downloads and calls add_subdirectory() , the lower-level population command allows you to retrieve files and stop there. Core Method: Manual Population