Liste des Groupes | Revenir à cl c |
David Brown <david.brown@hesbynett.no> writes:On 03/06/2024 11:00, Michael S wrote:On Sun, 2 Jun 2024 21:44:01 +0200
David Brown <david.brown@hesbynett.no> wrote:
C++ was viable for the kind of systems you were working with (clearly that is true, since you worked on an OS written in C++ at that time).Oh, there's little doubt that most publicly available RTOS kernels areI would disagree with this. The Chorus microkernel (Chorus Systemes,
in C, not C++. That does not mean C is in any way /better/ for the
task. There are multiple reasons for C being the language of choice here:
>
1. Most well-known RTOS kernels have a history stretching back to the
previous century. C++ was not nearly as viable an option at that time,
for a great many reasons.
later purchased by Sun) was started in the late 1980's and was
written in C++ (with a small set of assembler functions). This was
using Cfront (2.1 and later 3.0). I'm pretty sure it is still in
use. This was long before templates, exceptions or the standard library.
Many - I think most - RTOS's are linked as libraries, rather than separately linked applications.>Clearly one can use C interfaces from C++ code. And one can develop
2. If you write your kernel in C++, you pretty much have to use C++ for
the application code unless you also write a C API for it.
C++ wrapper around C-type functionality.
Our C++ kernels supported standard unix-style APIs between user
mode software and the kernel.
If you writeIf you write your kernel in _any_ lanaguage, you can use _any_ language
your kernel in C, you can use almost any language for the application code.
for the application code, or the kernel isn't much use to anyone.
Les messages affichés proviennent d'usenet.