Re: [NTLK] Using UnixNPI at 230400bps?

From: Victor Rehorst (
Date: Thu Mar 21 2002 - 21:10:26 EST

On Thu, 21 Mar 2002, Vladimir Alexiev wrote:

> Yes, however it doesn' t improve the speed for me. Here's the time
> required to transfer 172k, as reported by \bin\cygwin\time:
> 57600 49
> 115200 33.5
> 230400 33.5
> Using "start /realtime" improves the time by about 1s, but this
> measurement probably isn't very accurate.
> This makes 3.5kB/s for the slow speed and 5.1kB/s for the faster
> speeds. But 57600bps should give about 5.7kB/s (assuming 1 start and 1
> stop bit), so some speed is lost somewhere... (The time includes
> connection, handshake and finish-up, so I thought the transfer rate on
> bigger files will be better, but on a 1.5Mb file I got 5.6kB/s).

Hrm. I guess I put in the wrong flags into my termios.h, but it still
worked. *shrug* This is going to need some more investigating by

> In fact I can start ser230400 on Newton and "unixnpi -s 230400" on the
> PC and they connect fine. So I guess my cygwin libraries may be
> downgrading the speed. Time to upgrade my cygwin?

I don't think that would have anything to do with it.

> All the GUI I need is a shortcut like this on my desktop:
> unixnpi -s 230400 -d /dev/ttyS0 %1 %2 %3 %4 %5 %6 %7 %8 %9
> (optionally, prepend "cmd -c start /realtime ").

Hey, that's neat! I didn't know that Windows would add files as
parameters when you dropped something on a shortcut like that. Awesome.
Bye Bye NCU :)

> Please find below a little patch to make an error message more
> correct:
> If you run
> > myunixnpi.exe -s 115200 -d /dev/ttyNOSUCH whatever.pkg

Thanks, I've saved it. It'll be in the next release. In future, if you
could provide patches as a diff and post them on the SourceForge project
page, it would help me keep track of things.

Victor Rehorst - -
NewtonTalk list administrator -
Will help you with your list problems - as long as you're nice.

-- Read the List FAQ/Etiquette: Read the Newton FAQ: This is the NewtonTalk mailing list -

This archive was generated by hypermail 2.1.2 : Tue Apr 02 2002 - 14:03:22 EST