Mail Archives: djgpp/2006/03/09/08:30:51
Florian Liebig <FlorianLiebigEMD AT compuserve DOT de> wrote:
> Hans-Bernhard Broeker schrieb:
> > I've been meaning to ask you this before: why do you use such an
> > incredibly ancient version of GCC?
> because i've purchased an gnu c/c++ compiler a few years ago, a package that
> came on cd. (ISBN is :
> 3-8266-2714-8, price was 39.80 deutsche mark :^/ ).
> i've downloaded an update later, like from the pentium compiler group, that
> designed a cc1.exe, gcc.exe for the pentium processor.
Both that book and that download are excessively out of date. The PCG
has stopped to exist years ago. None of the persons who might be able
or willing to help you are still using such archaic tools. We've all
forgotten whatever details we knew about that version.
So do yourself a favour: update to the current version.
> > Similar question: why write it in C++, of all things?
> That is because I want to use OOP things, like classes with its automatic
> const./dest. calls. I like this.
I also like the comfort of riding in a nice Mercedes Benz S-class
limousine (well, hypothetically speaking ;-). That doesn't make it
the adequate vehicle for an off-road trip across the Himalaya or
through the Gobi desert, though. Choose your tools to match the job.
> That doesn't sounds good. So, there is no way round designing the kernel in
> pure ansi c code,
It can be done in mostly pure ANSI C. Doing it in C++ is, however, an
entirely different kettle of fish. It's strictly impossible to do it
in completely pure ANSI C, though. A C program is not like Baron
Muenchhausen --- it can't pull itself up out of the swamp by its own
boots. At least some parts of the startup code *have* to be written
in assembly, and you have quite a bit to learn about your CPU and the
internals of a working C program before you'll know what exactly those
parts are.
--
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.
- Raw text -