home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
kermit.columbia.edu
/
kermit.columbia.edu.tar
/
kermit.columbia.edu
/
newsgroups
/
misc.19980424-19980901
/
000120_news@newsmaster….columbia.edu _Fri May 22 16:51:43 1998.msg
< prev
next >
Wrap
Internet Message Format
|
1998-08-31
|
5KB
Return-Path: <news@newsmaster.cc.columbia.edu>
Received: from newsmaster.cc.columbia.edu (newsmaster.cc.columbia.edu [128.59.35.30])
by watsun.cc.columbia.edu (8.8.5/8.8.5) with ESMTP id QAA29012
for <kermit.misc@watsun.cc.columbia.edu>; Fri, 22 May 1998 16:51:41 -0400 (EDT)
Received: (from news@localhost)
by newsmaster.cc.columbia.edu (8.8.5/8.8.5) id QAA14045
for kermit.misc@watsun; Fri, 22 May 1998 16:51:41 -0400 (EDT)
Path: news.columbia.edu!panix!nntprelay.mathworks.com!newsfeed.internetmci.com!4.1.16.34!cpk-news-hub1.bbnplanet.com!wtn-news-feed2.bbnplanet.com!news.bbnplanet.com!netnews.jhuapl.edu!usenet
From: Skip Collins <collibf1@jhuapl.edu>
Newsgroups: comp.protocols.kermit.misc,comp.sys.hp48
Subject: Re: Kermit on the HP48 (Was: One-Way Transfer)
Date: 22 May 1998 16:47:51 -0400
Organization: Johns Hopkins University Applied Physics Lab, Laurel, MD, USA
Lines: 100
Sender: collibf1@COLLIBF1
Message-ID: <wk4syi58y0.fsf@jhuapl.edu>
References: <35646665.EBB3868B@theriver.com> <wkvhqye9g4.fsf@jhuapl.edu> <6k42pd$a3m$1@apakabar.cc.columbia.edu> <wk67iy1b8j.fsf@jhuapl.edu> <6k4ef6$g6p$1@apakabar.cc.columbia.edu>
NNTP-Posting-Host: collibf1-2.jhuapl.edu
X-Newsreader: Gnus v5.5/Emacs 20.2
Xref: news.columbia.edu comp.protocols.kermit.misc:8773 comp.sys.hp48:81334
fdc@watsun.cc.columbia.edu (Frank da Cruz) writes:
> Most Kermit programs prefix all control characters by default when sending a
> file. Kermit 95 is the exception, since most Windows 95 users demand "high
> performance". Kermit 95's default is to unprefix a fairly safe subset of
> control characters.
If I had to bet on what trips up most K95-HP48 connections I would wager it
is control-char unprefixing. I will try it with mskermit when I get a
chance. What command should I issue to mimic the K95 default?
> : I think the physical link is transparent to control characters. But perhaps
> : the hp48 kermit software assumes prefixing.
> :
> The cardinal rule of any communications protocol is "be conservative in
> what you send, be liberal in what you receive". The HP48 has it backwards.
> There is no prohibition in the protocol definition against sending bare
> control characters,
This seems to contradict "Kermit: A File Transfer Protocol", 1987 ed.,
which states on page 248, under the heading Encoding Summary: "Prefix
encoding for control characters is mandatory."
> : The serial port is full duplex. The infrared port is only half-duplex
> : because of optical feedback.
> :
> Does this apply to all models? I got my information about dead periods in
> the serial port from an HP engineer, circa 1990.
My information comes from HP and apparently applies to all HP48s. You
might want to have a look at
http://hp48.ml.org/programming/docs/ioguide.zip
Perhaps the dead time you mention is an artifact of a poor kermit
implementation. If you believe HP, it does not seem to be a limitation
of the hardware.
> Can you help clear this up? What is the deal? Text-mode transfers into
> the HP-48 are the ones that get progressively slower? But binary-mode
> transfers into the HP-48 are OK?
Both are affected.
> Is it true that incoming text-mode packets are parsed as HP-48 programs?
> So this means that only HP-48 programs may be sent in text mode, and any
> other text files sent to the HP-48 are likely to be rejected. One user
> reported that any text file containing a "-" character would be rejected
> for "Invalid syntax".
>
> This means that non-HP-48-program files must be sent to the HP-48 in binary
> mode, right?
The packets are parsed into an HP48 _object_, which might be a
program, or might just be a string variable. I think it tries to parse
the incoming stream as if it were being keyed in. So, [1 2 3], if sent
in text mode would become vector. If the syntax of the sent object
does not make sense as a keyed-in object, it becomes a string.
> set carrier-watch off ; (Does the HP-48 assert DCD? If so, use ON.)
It has no hardware flow control, if that's what you mean.
> set flow none ; (or Xon/Xoff)
Should be none.
> set control prefix all ; (Don't send bare control chars)
Until I see that unprefixing works, I should assume not.
> set receive timeout 99 ; (see Note 1 below...)
> set send pause 100 ; (see Note 2 below...)
I don't think these are needed, your notes notwithstanding. I very
rarely see packets retransmitted, which one might expect if there were
a serious turn-around problem.
> set file type text ; For HP-48 programs
Most programs people download to their calcs are compiled
binaries. They would be corrupted in text mode. Simple user-rpl
programs, such as one might key into the calc itself, can be
downloaded in text mode.
> set file type binary ; For all other forms of data
This is what I use almost exclusively. All the great freeware software
out there should be downloaded this way.
> Oh, by the way, another peculiarity I noticed... If the HP-48 is set up to
As long as you are cataloguing peculiarities, the HP48 sends packet
sequence numbers that are larger than 63. One can compensate by doing
mod63 before verifying. Is this right?
Another place to get more information is the hp48 faq:
http://www.engr.uvic.ca/~aschoorl/faq/48faq-6.html
> - Frank
Skip