I'm not familiar with HermitCore but the authors "...promise that this will make it easier to maintain and extend our kernel [originally written in C]." It'd be great if they could elaborate.
It may or may not be easier to "maintain and extend" a Rust implementation, but it will surely be even easier to find people to do those things if they keep it in C.
EDIT: Rephrased the sentence
I don't know about that, they teach C at pretty much every reputable University in america
That was my point - a lot more people know C
There's quantity, but what about quality? Whenever there's a vulnerability found in a C program, its authors are dismissed not real C programmers (since a real programmer would know to avoid buffer overflows and dangling pointers).
Apart from DJB, is there any other real C programmer? ;)
> Whenever there's a vulnerability found in a C program, its authors are dismissed not real C programmers (since a real programmer would know to avoid buffer overflows and dangling pointers).
This was the first time ever I've heard of such a thing.
And yet, I interview recent Comp Sci grads from upper tier University of California schools with 3.75 and above GPAs who try and tell me that a null pointer in a struct takes up zero memory. Then, when I ask them to correct themselves, I sometimes find out they have no idea how much data a pointer takes up.
Well, that is implementation defined, surely the only reasonable answer here is `sizeof(void * )` chars... where a `char` is probably 8 bits.
Or, they could ask if they are on a 32 bit or 64 bit system. They don't have to be absolutely pedantically correct. They just need to demonstrate they have a clue, as opposed to none.
I feel obliged to ask: are there any modern (post-1990) ISAs where a char _isn't_ 8 bits?