Re: zmodem

[llunch@knuth.cba.csuohio.edu (Jason Baker)]

> 
> The good news is that Zmodem 3.32 compiled under MiNTLIB 44 with only
> 1 change (addition of __default_mode__).  It generated 1 warning - whcih
> was a pointer type argument and this is a Zmodem problem.  3.32 is
> buggy though and sends to stderr when it should to stdout and vice versa.
> It also re-opens the terminal (here's the bug, by default whatever is
> connected to stderr - handle 2) with open() and dev/cua0 doesn't like
> this.   Once this fix has been done (open /dev/tty instead which is
> already in the zmodem source) then things should work fine, with one
> exception.   SZ is still broke (RZ is working with the BIOS AUX).  I
> think its a flow control problem.  Anyone have any ideas?   I couldn't
> find any lock calls or any mentioned of st_dev or st_rdev or any dev
> stuff at all, so 3.32 must be significantly different.  I compiled for
> posix, not bsd.
> 
	I just compiled zmodem 3.34 with mintlib pl 34.  The lock
calls are still in the V7 parts of rbsb.c.:
			ioctl(Tty, TIOCEXCL, 0);
and
		ioctl(Tty, TIOCNXCL, 0); 
The dev stuff does not seem to be there any more though.  Rz and sz
both work fine for me.  Incidentally, I am still using mint 1.10h3,
and sz is way faster than xyz -u.  With no serial fixes, xyz (the
version I have anyway) can upload at about 900cps, and sz at about
1860.  With modem0 installed it seems virtually impossible to slow sz
down.  Of course, sz is much nicer to other processes as well.

	Jason