[FreeVMS] Re: First compilation


Subject: [FreeVMS] Re: First compilation
From: BERTRAND JoŽl (joel.k.bertrand@free.fr)
Date: Wed Nov 21 2001 - 12:54:43 CET


Henry W. Miller wrote:

>
> ----- Original Message -----
> From: <roart@nvg.ntnu.no>
> To: <freevms@ml.free.fr>
> Sent: Wednesday, November 21, 2001 01:59
> Subject: [FreeVMS] Re: First compilation
>
>
>
>>On Tue, Nov 20, 2001 at 05:36:22PM +0100, BERTRAND JoŽl wrote:
>>
>>>BERTRAND JoŽl wrote:
>>>
>>>I have tried... Booting, and after 5 mn, the kernel is dead ;-)
>>>I'll put a full tarball on my web site.
>>>
>>>New plan :
>>>- remove the UTS_RELEASE and KERNELRELEASE from sources ;
>>>- clean the sources (remove KERNEL_VERSION and LINUX_VERSION_CODE tests)
>>>
> ;
>
>>>- find a logo to replace Tux ;
>>>
>>Shark?

        Shark ? What is Shark ?

>>>- modify Init becaus Init tests the number of kernel, and with a 0.0.1,
>>>I receive a kernel panic.
>>>
>>We should rather wait with such things, since the kernel changes pr date
>>
> is
>
>>only an approximately flat round-robin scheduler.
>>
>>We ought to get at least interrupts and IPLs in place before we do any
>>own numbering.
>>
>>-Roar Thronśs
>>
>>
>
> At first thought, I thought that we should integrate our numbering ASAP.
> But, upon further reflection, it occurs to me that there may be routines in
> the kernel that check the version upon compilation or execution.

        Yes, many routines. On the first kernel that I have archived (only to
find more volunteers, and I have found ;-) ), I have modified the
configuration script and patched the kernel to have ext3 support. I have
tried to modify the kernel version number, but it will be a big work
because the code contains a lot of #if/#endif structures based on three
numbers... And a lot of code is never compiled...

        In another hand, init panics when the kernel number is 0.0.1 (Kernel too
old ! Panic !). Thus, we shall need to fixe the kernel sources and the
init process to boot.

> Also, are we really wed to using Linux as a base? I thought that I had
> mentioned a couple of weeks back that there is another home-grown kernel
> group working on an "open source" OS compatible with Win NT.

        I think that NT is not a good OS, whitout any good memory protection and
without correct ressource management.

> It occurs
> to me that since it is pretty widely held that NT was in many ways a VMS
> clone since Dave Cutler was the architect of both systems.

        Dave Cutler has coded the protection access, not really the kernel. I
1996, I have bought a NT 4 source licence (I need this licence...), and
I have seen that NT was not written by Cutler, but by IBM engineers
which worked on OS/2 v1.3...

> Since this group
> is still pretty early on in it's development, there would not be a decade's
> worth
> of cruft to wade through to implement a VMS clone.
>
> I also ran across two interesting articles in the December 2001 issue of
> Dr. Dobb's Journal: two new "open source" systems, Menuet OS and NewOS.
> Of the two, it appears that NewOS would be the better of the two to
> implement
> a VMS clone upon.
>
> Does anyone have any idea what happened to the code and libraries that
> had already been accomplished by Rich Levitte's group? As I recall, they
> had several system libraries partially cloned, as well as a DCL compatible
> command processor.

        Rich Levitte and his group have only written a very little part of system
libraries. I don't believe that they have written any DCL compatible
interpreter.

        Regards,

        JKB

 

-- 
Liste de diffusion FreeVMS
Pour se dťsinscrire : mailto:freevms-request@ml.free.fr?subject=unsubscribe



This archive was generated by hypermail 2b25 : Wed Nov 21 2001 - 12:54:16 CET