Re: ideas

lilo (TaRDiS@mail.utexas.edu)
Wed, 8 May 1996 12:14:11 -0500 (CDT)


On Mon, 6 May 1996, Brian M Grunkemeyer wrote:

> Excerpts from internet.computing.linux-kernel: 5-May-96 Re: ideas by
> Alan Cox@lxorguk.ukuu.or
> > C++ code is currently slower and has buggier output than well written C (
> > especially as certain people bother to peek at the assembler output). Unless
> > the g++ walking volatile bug has finally been fixed its also not
> > usable. Finally g++ eats memory - forget building on an 8Mb machine with g++
>
> Ok, those are good temporary reasons to not use C++. Those problems
> should improve or eliminate themselves over time, when we might want to
> change our minds. But until then, how about making the kernel
> _compilable_ under C++? The advantages in the stronger type checking
> alone should be more than enough reason to periodically compile the
> kernel w/ g++ to make sure there aren't any subtle type errors. In a
> project like this where everyone's submitting patches randomly, it is a
> good safety measure. There's even a chance it could improve some code
> by pointing out which parameters weren't used in a function. While it
> doesn't guarantee good code, it does prevent sloppy code from causing
> problems.

If I recall correctly there's been one pass at this already. I suspect it
resulted in some improvement to g++ before it was broken off as
``currently-unworkable''.

It might be worthwhile to try this again, but I suspect the end of the 1.3.x
cycle is not the time to do it. Think about it as something worthwhile to
play with in 2.1.x. It's worth bringing up with Linus at that point....

lilo