While a lot of programs do compile and run, many executables have silent miscomplies. Many of the miscompiles don't reveal themselves until much later into the program's execution. Fixing this will require deeper usage of the compiled programs themselves.
C++ is a extremely large language and there is much to implement. The near term goal is to get the C++ standard library headers up and running.
With the worldwide memory shortage it has become important more software to reduce the amount of RAM they use. Aburiscript is no exception.
I am always looking for ways to make the compiler faster. However, it shouldn't come at the cost of code readability or maintainability.
The development of this compiler mainly took place on Apple Silicon. I am hoping to add support for aarch64 linux, either through a VM or a dedicated piece of Linux ARM hardware.
I do not have reliable access to an x64 machine at the moment. Support for x64 might come later but it is not a big priority.
It seems possible to create a non-optimizing backend that beats the existing compilers on compile time. This would obviously be a lot of work. I used LLVM as a backend because the frontend work is also difficult. Using a known correct backend it makes it easier to isolate bugs in the smaller code surface.