[Coco] I should know better

[William Schaub]

Thanks for verifying that this is the case. I read that too and thought 
it would be useful to take advantage of that somehow.
I'm hoping that maybe it does indeed look in memory somehow but that 
more than just having the raw module image in a memory location of your 
choosing would be needed, like say also putting a pointer to the module 
at a certain location or having extra header information poked into 
memory with the image or something.




Willard Goosey wrote:
> ...than to believe the Tandy OS-9 manuals.
>
> Just from reading the docs, I was under the impression that, on boot,
> OS-9 scanned through all of memory and linked in any modules it
> found.  
>
> Should have known it didn't, at least under LII.  That would take a
> long time.
>
> But I was wondering if it actually did, as I can see certain useful
> applications for this.  So I tested it.
>
> My Setup:  6809 512K CoCo3, NitrOS-9 3.2.6
>
> First, I looked over the output of mmap, and chose block 0x30, as one
> unlikely to get clobbered during the boot process.
>
> Used the rsdos command to copy an executable OS-9 module to a DECB
> disk.  (cputype, actually.  I wanted something small in case I had to
> do this by hand.)  
>
> Rebooted to DECB.
>
> Modified a BASIC program to read the file and LPOKE it into memory
> starting at 0x600000.  Another program to LPEEK and print the results
> to verify the module is there.
>
> Loaded cputtype in at 0x60000.  Verified it was intact.
>
> Booted NitrOS-9.
>
> Ran mdir.  No cputype.  Link couldn't find it either.
>
> Rebooted to DECB.  cputype still present in block 0x30.
>
> Looked through my disks and found a stock OS-9LII boot disk.  Booted
> that.  No cputype.
>
> So, on boot, OS-9LII does NOT scan through all of memory looking for
> valid modules.
>
> Willard
>