inline visitAstNodes() calls

[firewave]

See https://blog.demofox.org/2015/02/25/avoiding-the-performance-hazzards-of-stdfunction/.

Testing with cli/threadexecutor.cpp:
GCC 11 656,137,737 -> 582,567,111
Clang 13 605,911,420 -> 556,949,984

[firewave]

The inlining seems to behave differently with GCC since I still see calls to the function in question. There's also still a lot of std::function related overhead. Probably applies to #4146 as well. This will require some serious looking into to get it all the way like with Clang.

[pfultz2]

I think its better to use a template parameter instead of std::function wherever possible. Not only will we have better performance, but its easier to step through with the debugger.

[firewave]

I think its better to use a template parameter instead of std::function wherever possible. Not only will we have better performance, but its easier to step through with the debugger.

In the profiler it badly messed things up since it is being inlined. It's really annoying. But you still might get to the source of things.

In GCC somehow this doesn't work as expected. I haven't filed as much reports with GCC as Clang yet and this really needs some looking into but that will probably take quite a while for me to get to...