FYI one of the lesser known features of cmake is that it can invoke your build tool as well:
cmake ..
cmake --build .
This smartly invokes make/ninja/msbuild or whatever directly from the command line; very useful on windows machines or to have 'universal' cmake runners for CI that touches multiple platforms.
(Also, tutorials; `cmake .. && cmake --build .` should smartly determine the correct compiler for the system and build it; you don't need the -G "FOO" stuff unless you're specifically doing something fancy)
To help improve my OpenGL/graphics knowledge, I've started going through GPU Gems ( https://a248.e.akamai.net/f/248/10/10/http.developer.nvidia.... ) and building my own examples of the concepts in each chapter. I was very happy to find Glitter recently, as it made the obnoxious hurdle of setting up all the dependencies unnecessary.
Also, setting up Glitter was what led me to finally "getting" what CMake was for and how it fit into a development pipeline.
Is there a particular reason you choose to explicitly add CXX flags ('-std=c++11') over CMake's compiler feature support? The later is, in principle, more portable and future proof than hardcoding the flags.
This was asked in an issue earlier today. It's a combination of 1) I didn't even know such flags existed! and 2) I had some compatibility issues with older versions of CMake on my test machines.
By "developer" is this from the perspective of say, a company looking to use Glitter as a starting point for something?
Glitter started off as a quickstart for students taking Computer Graphics related classes at my school, not as the foundation for writing commercial game engines.
Was there something in mind that you feel like is missing? I'm happy to add to the readme.
(Also, tutorials; `cmake .. && cmake --build .` should smartly determine the correct compiler for the system and build it; you don't need the -G "FOO" stuff unless you're specifically doing something fancy)