C++ .gitignore template

Everything the C template covers applies here too: object files, static and shared libraries, executables and debug symbols. C++ adds nothing new to that list, because the difference between the two languages disappears by the time the linker runs. What C++ projects add in practice is a build system, and that is where the extra rules come from.

CMakeCache.txt is the single most important addition. It records every configuration variable resolved during the configure step, including the absolute path to the compiler and to the source tree on the machine that ran it. Copying it to another machine produces confusing failures that look like compiler bugs, which is why a committed cache is worse than no cache at all.

CMakeFiles is the directory of generated build rules and compiler feature probes that sits beside the cache, and cmake_install.cmake is the generated install script. cmake-build-* matches the directory naming convention that CLion uses for its configurations, so a project opened in that editor will produce cmake-build-debug and cmake-build-release without anyone asking for them.

The Makefile rule is the one to think about. It exists because CMake generates a Makefile in the build directory, and generated files should not be tracked. But plenty of C++ projects also keep a hand written top level Makefile as a convenience wrapper, and this rule ignores that file too, with no warning. If your Makefile is something you wrote, either remove the line or add a negation for the exact path.

The safer arrangement, and the one most CMake projects settle on, is to configure out of source into a directory called build and ignore that directory wholesale. Then none of the individual CMake rules matter, and a hand written Makefile at the root stays tracked because nothing is trying to match it by name.