Extract SafeBinaryMutex to separate header

[vittorioromeo]

This change simply extracts 'SafeBinaryMutex' from 'mutex.h' to 'safe_binary_mutex.h', in an effort to reduce the compilation speed impact of including mutex.h.

Similarly in nature to #87871, this is a pure refactoring PR that shouldn't affect the behavior or Godot in any way or form.

[AThousandShips]

Can you show some data to show this reduces compile times?

[vittorioromeo]

Can you show some data to show this reduces compile times?

I've benchmarked the compilation of the entire engine using ClangBuildAnalyzer and -ftime-trace. Here's the original report:

https://gist.github.com/vittorioromeo/fa153c494ce05da08c988a27130a79e3

There are a few heavy-hitters here and there, and not everything is easy to optimize, so I am starting with some low-hanging fruits. The "expensive headers" section of the report shows that mutex.h is included quite often and takes a considerable amount of time.

Reducing the amount of code that mutex.h contains is a way to reduce the impact of including that header. Furthermore, splitting a specialized components (SafeBinaryMutex) that is only used in one place to its own header seems like good engineering practice to me regardless of what the impact on compilation time might be.

[AThousandShips]

Can you show after this change as well? Showing the problem without indicating this is a solution would be missing the important part IMO, i.e. the showing it improves compile time

[vittorioromeo]

Can you show after this change as well? Showing the problem without indicating this is a solution would be missing the important part IMO, i.e. the showing it improves compile time

https://gist.github.com/vittorioromeo/f59a1f1e312beb603977256059681717

# BEFORE
779723 ms: core/os/mutex.h (included 1104 times, avg 706 ms), included via:

# AFTER
758503 ms: core/os/mutex.h (included 1110 times, avg 683 ms), included via:

[Geometror]

Looks fine to me, you just need to run clang-format or sort the includes manually by lexicographical order.

[vittorioromeo]

@Geometror: should be good to go now! :)
Can you also please look at #87871 ?

[akien-mga]

Makes sense to me.

[akien-mga]

Thanks!