[NTLK] NEWT/0 set-aref bug

Matthias Melcher mm at matthiasm.com
Tue Dec 23 22:46:22 EST 2014

> On Dec 23, 2014, at 11:29 PM, Morgan Aldridge <morgant at makkintosshu.com> wrote:
> On Tue, Dec 23, 2014 at 4:17 PM, Jake Bordens <jake at allaboutjake.com> wrote:
>>> I wonder if we can finally find some way to connect NCX to Einstein,
>>> preferably via a simulated Ethernet connection That would make
>>> synchronization possible for all Einstein devices on all platforms.
>> I've been thinking about this as well.  I was thinking along the lines of
>> retargeting the methods on the SerialChip class so that Einstein could
>> read/write to a pipe on the host file system that other applications could
>> access.  Maybe even Basilisk could be configured to read/write to that
>> pipe and allow one emulated system to connect to another.
>> I had hoped it would be simple-- a getch/putch interface that could be
>> easily retargeted.  Looking at the symbols, I think its more complicated
>> than I thought.  There is a registry of serial devices
>> (SerialChipRegistry) which I assume is for adding/removing PCMCIA devices,
>> an possibly the built in ports.  The SerialChip class is a bit more
>> complicated than just a source/sync of bytes.  In the end, the idea is
>> probably above my understanding of how NewtonOS handles the serial ports.
> I know Eckhart Köppen had considered doing it this way once as well.
> That would certainly be useful for NTK in Basilisk or tntk.

I have spent a few days in that code, but I did not find an easy solution to emulate a serial port. Maybe I looked in the wrong places, but I found it quite difficult. NewtonOS seems to use a DMA channel for this, making all the data transfer completely asynchronous.

More information about the NewtonTalk mailing list