Re: An OS in Ada, why not RTEMS ?

[James Ross <rem.jr@rem.webross.com>]

>> We need those machine language bytes that are in the .lst file with
>> the proper address of the _system__storage_elements__to_address
>> subroutine along with the subroutine ml code itself in a .bin file.
>> Not an ELF, not a PE, not a .o, not a .lst, But a raw binary file.
>>
>> CAN WE DO THAT?
>
>Yes, we can, several people have done this already (me included). Actually,
>what's wrong with ELF? GRUB(http://www.gnu.org/software/grub/) can load it.
>But if you still want a .bin, I believe it is possible to get it using GNU
>binutils. However, the hard question seems to be "what now?"

My idea *was* to just support a raw binary executable image that not
only the kernel would be in, but other programs that could be loaded
and executed by the kernel.  (Perhaps I needed more info on this
issue,  but I *was* certain it would be possible).  If I were willing
to dig into the details of ELF to support loading it right away, then
I might as well chose PE ... since I prefer to use GNAT from (my main
computer) Windows XP.  (And of course GNAT on Windows cannot create
ELFs.)  I still don't use Linux as my main computer - but that is a
whole other issue :)   

As for the initial boot process and the ELF issues -- understand that
what I have been saying is how I *was* going to tackle the process if
it were up to me.  I have all but given up at this point.  If I were
to help someone else, then the design would be whatever basically
exists and works for that person(s) ...  I would be willing to work on
a Linux box (I am not allergic to it!) and support whatever version of
an Ada OS emerges ...

JR