home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
kermit.columbia.edu
/
kermit.columbia.edu.tar
/
kermit.columbia.edu
/
e
/
misc.save
/
000143_fdc@panix.com_Thu Dec 20 16:44:53 2007.msg
< prev
next >
Wrap
Internet Message Format
|
2011-07-03
|
4KB
Path: reader1.panix.com!panix!not-for-mail
From: Frank da Cruz <fdc@panix.com>
Newsgroups: comp.protocols.kermit.misc
Subject: Re: Screen Capture
Date: Thu, 20 Dec 2007 21:44:43 +0000 (UTC)
Organization: PANIX Public Access Internet and UNIX, NYC
Lines: 58
Message-ID: <slrnfmloib.rsq.fdc@panix2.panix.com>
References: <h654j.66697$RX.26340@newssvr11.news.prodigy.net> <13l6v29jqcmhl2e@corp.supernews.com> <g3J7j.77732$YL5.57974@newssvr29.news.prodigy.net> <13m3fqgakmrkc05@corp.supernews.com> <XWm8j.7512$AR7.2068@nlpi070.nbdc.sbc.com> <slrnfm5ghr.8om.fdc@panix3.panix.com> <80oaj.374$6%.86@nlpi061.nbdc.sbc.com>
Reply-To: fdc@columbia.edu
NNTP-Posting-Host: panix2.panix.com
X-Trace: reader1.panix.com 1198187083 9750 166.84.1.2 (20 Dec 2007 21:44:43 GMT)
X-Complaints-To: abuse@panix.com
NNTP-Posting-Date: Thu, 20 Dec 2007 21:44:43 +0000 (UTC)
User-Agent: slrn/0.9.8.0 (NetBSD)
Xref: panix comp.protocols.kermit.misc:15725
On 2007-12-20, 2damn <2damn@nospam.com> wrote:
: On Fri, 14 Dec 2007 17:49:47 +0000, Frank da Cruz wrote:
:> C-Kermit can handle text-mode dialogs just fine, and for that matter
:> also menu-driven dialogs, if you can do this without specific reference
:> to screen positions and coordinates. For example, the host application
:> paints a manu and at the bottom it prints "Choice?" and the user is
:> supposed to enter the number (or whatever) of the desired menu item.
:> Well, if "Choice?" is the last thing that the host sends (which is
:> something you can tell from a session log), then:
:>
:> clear input
:> input 10 Choice?
:> if failure (do something)
:> output 3\13
:
: I have the input, if failure, output routine going okay. I am currently
: using a generic FAIL message and will add better notices where needed.
:
: The trouble is that each feature is not exactally the same at all times.
: As I said I think kermit can deal with this. At the very least, well
: enough to send unsolvable items to an error report.
: ...
: The system accepts commands just fine so long as I can compensate for
: some differences. The issue with garbled feedback applies only to "screen
: grabs". Using the kermit session record some "screens" just have
: "interuptions" with serial communication data and some seem completely
: broken and unrecognizable. I think the kermit session logging is not the
: way for me to correctly capture the feedback from the system.
:
Session logging has nothing to do with scripting. If your script is
supposed to be interacting with a remote menu or whatever, then the script
uses INPUT to read incoming material and reacts to it with OUTPUT.
But any form of scripting relies on the messages, prompts, menus, etc,
being received intact. If you can't predict what is going to come, how can
you reasonably write a script to react to it?
If you have a network connection, there should be no garbling because
network protocols take care of error correction and flow control. If you
have a serial port or modem and you're getting garbage, you need to fix
the garbage. If it's a modem, you need to be using an error-detecting and
-correcting protocol between the two modems, and then (mode or direct
connection) you need to have the most effective possible form of flow
control enabled, which normally would be RTS/CTS (a.k.a. "hardware flow
control").
: This week I have individual items running through 2 processes from a list
: thanks to the online kermit documentation. Adding an option for a third
: process and a useable kermit or bash menu will bring it up to snuff as a
: beta usable to others.
:
: Thanks for all the info so far.
:
Sure. And remember, the more specifics you give about your connection,
the behavior of the applicaiton you're scripting, etc, the better we can
help you.
- Frank