Re: Changing from L4 to something else...

[Jonathan S. Shapiro]

On Fri, 2005-10-28 at 17:38 +0200, Alfred M. Szmidt wrote:
>    If certain people wrote one line of code for every hundred lines of
>    argumentation email, Hurd would be finished by now.
> 
> I just did, did you?  Infact, I have written aprox 1000 lines of code
> over the past two days.

Here is a more direct answer, and I am only answering because it may be
of actual interest what I have been up to.

I did manage to get about 80 lines of code out, and in this particular
case I think that almost all of it will be retained.

Anybody who knows the Pentium will know that initializing this piece of
crap properly is by far the hardest part of building a kernel for it.
Much of yesterday's code was the kind where you add one assembly
instruction, recompile, retest, and then go on -- because the main
diagnostic of failure is that the machine reboots.

In order to get that code built, I had to work through how kernel-layer
per-CPU storage would be handled, which is actually a non-trivial design
problem.

So: not much progress in terms of stuff that anybody will ever see, but
a major, nasty piece of really bitchy code is now fairly stable. Once I
get the IDT set up properly things should pick up a bit.


shap