It's ok to re-iterate over the Cmake cache?

Question

I wasn't sure how to set the title for this.

I basically have a code base that I build with cmake; now I would like to know if it's safe to invoke cmake itself on the same cache more than 1 time.

For example in the dir /path I'm building a cache with

cmake -DVAR=VALUE < more flags >

and then in a second step I invoke cmake again in the same directory /path with another set of flags

cmake -DANOTHER_VAR=ANOTHER_VALUE < more flags >

at this point there are some guarantees that what is inside /path is the result of both commands ? It depends on how the CMakeLists is written ?

My building system that I use over cmake could be simplified a little thanks to this steps, that's why I'm asking.

Answer 1

Yes, it's OK to do this. Actually, when you change something in your CMakeLists.txt's and then run make in build folder, CMake invokes cmake . there to regenerate the cache.

Beside that, -DVAR=VALUE is an intended way for setting vars in the generated cache, as well as -UVAR.

The only exception is changing vars like CMAKE_C_COMPILER or CMAKE_CXX_COMPILER. If you alter them with cmake -DCMAKE_C_COMPILER=newcc ., CMake would wipe your cache and generate it from scratch.