[Coco] NitrOS9 and 6309 code

[L. Curtis Boyle]

I think we should make a unique error code that explicitly states that a 6309 module can not run on a 6809 system. We have a fair number of open spaces in the error codes to put such an error code in. “Bad module header” can mean corrupt modules, etc. as well, so I would prefer to differentiate it.


L. Curtis Boyle
curtisboyle at sasktel.net



> On May 31, 2019, at 7:43 PM, Allen Huffman <alsplace at pobox.com> wrote:
> 
>> On May 31, 2019, at 8:21 PM, L. Curtis Boyle <curtisboyle at sasktel.net> wrote:
>> 
>> And it should report an error unique to that situation, so that the user immediately knows what is wrong. Not “Bad module header”, which covers multiple circumstances.
> 
> 
> We can mirror what modern OS-9 does… If you try to load a 68K binary on a PowerPC machine, you get an error code. Let me see if I can figure what that is.
> 
> 
> /sd02/USR/ALLENH $load oscom_386
> load: can't load "oscom_386" - Error #000:205
> 
> /sd02/USR/ALLENH $load oscom_ppc
> load: can't load "oscom_ppc" - Error #000:205
> 
> Error 205 is:
> 
> 000:205 	E$BMID 	Bad module ID
> 
> This means that the Module ID at the start is bad or unrecognized. I guess this isn’t totally accurate, since the module type would be the same, and only the sub type is changed?
> 
> 6809 and 6309 are different CPU types, so maybe it should be a different module ID value, like the other OS-9s do it. 
> 
> Actually, I wonder what happens if you try to load a 68040 module on a 68000. That’s what we need to look up. I have no way to do that currently.
> 
> Maybe a test with X86 is possible? Compile something targeting Pentium and see if it loads on x86?
> 
> 		— Allen
> 
> 
> -- 
> Coco mailing list
> Coco at maltedmedia.com
> https://pairlist5.pair.net/mailman/listinfo/coco
>