[1]The Columbia Crown The Kermit Project | Columbia University
612 West 115th Street, New York NY 10025 USA o [2]kermit@columbia.edu
...since 1981
[3]Home [4]Kermit 95 [5]C-Kermit [6]Scripts [7]Current [8]New [9]FAQ
[10]Support
C-Kermit 8.0 Update Notes
[ [11]Contents ] [ [12]C-Kermit ] [ [13]Kermit Home ]
Second Supplement to [14]Using C-Kermit, 2nd Edition
For C-Kermit 8.0
As of C-Kermit version: 8.0.211
Date of C-Kermit release: 10 April 2003
This file last updated: Mon Sep 13 08:52:36 2010
* IF YOU ARE READING A PLAIN-TEXT version of this document, note that
it is a plain-text dump of a Web page. You can visit the original
(and possibly more up-to-date) Web page here:
[15]http://www.columbia.edu/kermit/ckermit80.html
* If you are reading the HTML version of this file with a GUI Web
browser, the features added since C-Kermit 8.0.201 are shown in red
if your browser and monitor permit. Features that were new to
versions 8.0.200 and 201 are in black.
Authors: Frank da Cruz and Christine M. Gianone
Address: The Kermit Project
Columbia University
612 West 115th Street
New York NY 10025-7799
USA
Fax: +1 (212) 662-6442
E-Mail: [16]kermit-support@columbia.edu
Web: [17]http://www.columbia.edu/kermit/
Or: [18]http://www.kermit-project.org/
Or: [19]http://www.columbia.nyc.ny.us/kermit/
NOTICES
This document:
Copyright © 1997, 2002, Frank da Cruz and Christine M. Gianone.
All rights reserved.
Kermit 95:
Copyright © 1995, 2002, Trustees of Columbia University in the
City of New York. All rights reserved.
C-Kermit:
Copyright © 1985, 2002,
Trustees of Columbia University in the City of New York. All
rights reserved. See the C-Kermit [20]COPYING.TXT file or the
copyright text in the [21]ckcmai.c module for disclaimer and
permissions.
When Kerberos(TM) and/or SRP(TM) (Secure Remote Password) and/or
SSL/TLS protocol are included:
Portions Copyright © 1990, Massachusetts Institute of
Technology.
Portions Copyright © 1991, 1993 Regents of the University of
California.
Portions Copyright © 1991, 1992, 1993, 1994, 1995 by AT&T.
Portions Copyright © 1997, Stanford University.
Portions Copyright © 1995-1997, Eric Young <eay@cryptosoft.com>.
For the full text of the third-party copyright notices, see
[22]Appendix V.
WHAT IS IN THIS FILE
This file lists changes made to C-Kermit since version 7.0 was released
in January 2000. Use this file as a supplement to:
* The second edition of [23]Using C-Kermit; and:
* The [24]C-Kermit 7.0 Update Notes. Also available in plain-text
form as [25]ckermit70.txt.
until the third edition of Using C-Kermit is published. We apologize
for the scattered documentation and will consolidate it when we are
able.
ADDITIONAL FILES
Several other files accompany this new Kermit release:
[26]ckututor.html
C-Kermit Tutorial (for Unix). Also distributed in Nroff form as
[27]ckuker.nr, the Unix C-Kermit manual page.
[28]security.htm
Discussion of Kermit's new authentication and encryption
features, updated for C-Kermit 8.0.
[29]telnet.htm
Detailed documentation of Kermit's Telnet client, updated for
C-Kermit 8.0.
[30]ftpscripts.html
Tutorial: Writing FTP automation scripts
[31]ckcbwr.html
Platform-independent C-Kermit hints and tips. Also distributed
in plain text form as [32]ckcbwr.txt
[33]ckubwr.html
Unix-specific C-Kermit hints and tips. Also distributed in plain
text form as [34]ckubwr.txt.
[35]ckvbwr.html
VMS-specific C-Kermit hints and tips. Also distributed in plain
text form as [36]ckvbwr.txt.
[37]ckuins.html
Unix C-Kermit installation instructions. Also distributed in
plain text form as [38]ckuins.txt.
[39]ckvins.html
VMS C-Kermit installation instructions. Also distributed in
plain text form as [40]ckvins.txt.
[41]ckccfg.html
Compile-time configuration options. Also distributed in plain
text form as [42]ckccfg.txt.
[43]ckcplm.html
C-Kermit Program Logic Manual. Also distributed in plain text
form as [44]ckcplm.txt.
[45]iksd.html
Internet Kermit Service Administrators Guide for Unix.
[46]skermit.html
C-Kermit as an SSH Subsystem (SFTP server replacement).
[ [47]Top ] [ [48]C-Kermit ] [ [49]Kermit Home ]
CONTENTS
[50]0. WHAT'S NEW
[51]1. FIXES SINCE VERSION 7.0.196
[52]2. SSH AND HTTP
[53]2.1. SSH Connections
[54]2.2. HTTP Connections
[55]2.2.1. HTTP Command Switches
[56]2.2.2. HTTP Action Commands
[57]2.2.3. HTTP Headers
[58]2.2.4. Secure HTTP Connections
[59]2.2.5. HTTP Variables
[60]2.2.6. The HTTP Command-Line Personality
[61]3. THE BUILT-IN FTP CLIENT
[62]3.1. Making and Managing FTP Connections
[63]3.1.1. Kermit Command-Line Options for FTP
[64]3.1.2. The FTP Command-Line Personality
[65]3.1.3. The FTP URL Interpreter
[66]3.1.4. Interactive FTP Session Establishment
[67]3.2. Making Secure FTP Connections
[68]3.3. Setting FTP Preferences
[69]3.4. Managing Directories and Files
[70]3.5. Uploading Files With FTP
[71]3.5.1. FTP PUT Switches
[72]3.5.2. Update Mode
[73]3.5.3. Recovery
[74]3.6. Downloading Files With FTP
[75]3.6.1. FTP GET Switches
[76]3.6.2. Filename Collisions
[77]3.6.3. Recovery
[78]3.7. Translating Character Sets
[79]3.7.1. Character Sets and Uploading
[80]3.7.2. Character Sets and Downloading
[81]3.8. FTP Command Shortcuts
[82]3.9. Dual Sessions
[83]3.10. Automating FTP Sessions
[84]3.10.1. FTP-Specific Variables and Functions
[85]3.10.2. Examples
[86]3.10.3. Automating Secure FTP Connections
[87]3.11. Advanced FTP Protocol Features [88]4. FILE SCANNING
[89]5. FILE AND DIRECTORY NAMES CONTAINING SPACES
[90]6. OTHER COMMAND PARSING IMPROVEMENTS
[91]6.1. Grouping Macro Arguments
[92]6.2. Directory and File Name Completion
[93]6.3. Passing Arguments to Command Files
[94]6.4. More-Prompting
[95]6.5. Commas in Macro Definitions
[96]6.6. Arrow Keys
[97]7. NEW COMMANDS AND SWITCHES
[98]8. SCRIPTING IMPROVEMENTS
[99]8.1. Performance and Debugging
[100]8.2. Using Macros as Numeric Variables
[101]8.3. New IF Conditions
[102]8.4. The ON_UNKNOWN_COMMAND and ON_CD Macros
[103]8.5. The SHOW MACRO Command
[104]8.6. Arrays
[105]8.7. New or Improved Built-in Variables and Functions
[106]8.8. The RETURN and END Commands
[107]8.9. UNDEFINing Groups of Variables
[108]8.10. The INPUT and MINPUT Commands
[109]8.11. Learned Scripts
[110]8.12. Pattern Matching
[111]8.13. Dates and Times
[112]8.14. Trapping Keyboard Interruption
[113]9. S-EXPRESSIONS
[114]9.1. What is an S-Expression?
[115]9.2. Integer and Floating-Point-Arithmetic
[116]9.3. How to Use S-Expressions
[117]9.4. Summary of Built-in Constants and Operators
[118]9.5. Variables
[119]9.6. Assignments and Scope
[120]9.7. Conditional Expressions
[121]9.8. Extensibility
[122]9.9. Examples
[123]9.10. Differences from Algebraic Notation
[124]9.11.Differences from Lisp
[125]10. FILE TRANSFER
[126]11. MODEMS AND DIALING
[127]12. TERMINAL CONNECTION
[128]13. CHARACTER SETS
[129]14. DIALOUT FROM TELNET TERMINAL SERVERS
[130]15. COPING WITH BROKEN KERMIT PARTNERS
[131]16. NEW COMMAND-LINE OPTIONS
[132]17. LOGS
[ [133]Top ] [ [134]C-Kermit ] [ [135]Kermit Home ]
0. WHAT'S NEW
The Initialization and Customization Files
C-Kermit 8.0 now supports specification of the initialization
file name (path) in an environment variable, CKERMIT_INI. It
also relies far less than before on the initialization for
functioning. See [136]Section 5 of the Unix C-Kermit
[137]installation instructions for details. As of version
8.0.201, C-Kermit also executes your customization file (if you
have one) even if the initialization file was not found.
Previously, the customization file was executed by a TAKE
command in the initialization file (and it still is, if an
initialization is found).
Incompatible Changes
As always, we do our best to avoid changes that break existing
scripts. However, C-Kermit 8.0 does include a rather pervasive
syntax change that might alter the behavior of scripts that
depend on the previous behavior. As described in [138]Section 5,
C-Kermit now accepts doublequotes in most contexts where you
previously had to use braces to group multiple words into a
single field, or to force inclusion of leading or trailing
blanks. Most noticeably, in C-Kermit 7.0 and earlier:
echo {this is a string}
would print:
this is a string
whereas:
echo "this is a string"
printed:
"this is a string"
In C-Kermit 8.0, both print:
this is a string
To force the doublequotes to be treated as part of the string,
use either of the following forms:
echo {"this is a string"}
echo ""this is a string""
Similarly, to force braces to be treated as part of the string:
echo "{this is a string}"
echo {{this is a string}}
Other incompatibilities:
1. Using the SET HOST command to make HTTP connections is no
longer supported. Instead, use the new HTTP OPEN command,
described in [139]Section 2.2.
C-Kermit 7.1 Alpha.01 (8 December 2000)
Its major new features are those listed in the [140]Table of
Contents: the FTP client, file scanning, command parsing and
scripting improvements, S-Expressions, and support for the
Telnet Com Port Option, plus wider availability of the Kerberos,
SSL/TLS, and SRP security options for secure Internet
connections.
C-Kermit 7.1.199 Alpha.02 (4 January 2001)
+ C-Kermit now accepts [141]FTP, TELNET, and IKSD URLs as its
first command-line argument.
+ Character-set translation added to the FTP client for
[142]filenames.
+ Optional [143]setting of date of incoming files by FTP [M]GET
from the server date.
+ [144]FTP CHECK filename added to let FTP client check the
existence of a file on the server.
+ [145]FTP GET /NAMELIST:filename added to get list of server
filenames into a local file.
+ [146]FTP [M]PUT /SERVER-RENAME:template added to make server
rename a file as indicated by the template after it has
arrived completely.
+ FTP [M]GET /SERVER-RENAME:template added to make server rename
a file as indicated by the template after it has been sent
completely.
+ FTP [147]VDIRECTORY added for getting verbose directory
listings from TOPS-20.
+ [148]FTP TYPE TENEX added for transferring 8-bit binary files
with PDP-10s.
+ Added [149]automatic text/binary mode switching for FTP
[M]GET, based on filename patterns (e.g. *.zip, *.gz, *.exe
are binary; *.txt, *.c are text).
+ [150]SET SEND I-PACKETS OFF added for coping with Kermit
servers that do not support I packets.
+ A new option was added to [151]\fword() and \fsplit() for
parsing comma-separated lists that might contain empty
elements.
+ Bug fixes including:
o {} or "" could not be used as expected to represent the
empty string.
o ,- on a line by itself in a macro definition caused
subsequent statements to be skipped.
o FTP [M]GET didn't work right if path segments were
included in the filespec.
o FTP MGET, if interrupted, did not clear its file list.
o Various problems with FTP PUT /AS-NAME that nobody
noticed.
o Some FTP messages and displays interfered with each
other.
o Parsing of YESTERDAY, TODAY, and TOMORROW in date-time
fields was broken.
o Automatic old-to-new dialing directory format conversion
was broken on VMS.
o Various source-code portability problems fixed.
+ Improvement of various HELP and SHOW messages.
C-Kermit 7.1.199 Alpha.04 (1 April 2001)
+ Big changes:
o Changed default modem type from NONE to GENERIC.
o Generic dialing now sends no init string at all.
o Changed default terminal bytesize from 7 to 8.
+ New features:
o SET SESSION-LOG TIMESTAMPED-TEXT for timestamped session
log.
+ New modem types:
o Conexant modem family
o Lucent VENUS chipset
o PCTel V.90 chipset
o Zoom V.90
o Zoom V.92
+ FTP client:
o FTP OPEN /PASSIVE and /ACTIVE switches added.
o Now works with servers that that don't include path in
NLST response.
o Fixed SEND /RECURSIVE not to follow symlinks (UNIX).
o SET FTP VERBOSE-MODE default is now OFF instead of ON.
+ Kermit protocol:
o Fixed what I hope is the last "Receive window full"
error.
o SET PREFIXING or SET CONTROL PREFIX now automatically
sets CLEARCHANNEL OFF.
o Fixed incorrect report of number of files transferred at
end of transfer.
o Fixed SEND /RECURSIVE not to follow symlinks (UNIX).
+ UNIX:
o HTTP and shadow passwords enabled for SCO 5.0.6.
o Even with SET FILENAMES CONVERTED, spaces were still
accepted in incoming filenames; now they are converted to
underscores.
o Added support for compile-time mktemp()/mkstemp()
selection.
+ VMS:
o Session-log format for scripted sessions fixed.
+ Scripting:
o Fixed \frdir() not to follow symlinks (UNIX).
o Fixed \fday() not to dump core for dates prior to 17 Mar
1858.
+ General:
o "Closing blah..." message upon exit could not be
suppressed.
o Added /PAGE and /NOPAGE to DELETE switches.
o Added GO response for DELETE /ASK (delete all the rest
without asking).
o Added GO response to "more?" prompt (for multi-page
screen output).
o Updated HELP texts.
C-Kermit 7.1.199 Beta.01 (10 May 2001)
+ FTP client verbosity adjustments.
+ Bug with generic modem dialing pausing several secs fixed.
+ SET HOST /USER:, SET LOGIN USERID, etc, fixed when given no
user ID.
+ A couple \v(dm_blah) dial modifier variables added.
+ "--version" command-line switch added.
+ Fixed NetBSD serial-port DTR handling.
+ Lots of syntax cleanups for Flexelint and gcc -Wall.
+ Fixed modem-type aliases to not take precedence over real
names.
+ Fixed funny treatment of doublequotes by ECHO command.
+ Enabled SET SESSION-LOG for VMS and other non-UNIX platforms.
+ Fixed changing direction in command history buffer.
+ Fixed handling of IKSD URLs.
+ Made sure DELETE prints a message if it got any errors.
C-Kermit 8.0.200 Beta.02 (28 June 2001)
+ Major version number increased from 7 to 8.
+ [152]SSH command.
+ More-consistent Kermit protocol defaults.
+ CONNECT idle timeout and action selection.
+ CONNECT status variable.
+ A way to allocate more space for filename lists.
+ Pseudoterminal handler fixed for late-model Linuxes.
+ Command-line option -dd for timestamped debug log.
+ Download directory now works for external protocols too.
+ GREP /COUNT:variable.
+ SET ATTRIBUTE RECORD-FORMAT { OFF, ON }.
+ Bug fixes.
C-Kermit 8.0.200 Beta.03 (9 Sep 2001)
+ [153]HTTP 1.1 connections and scripting
+ [154]ON_CTRLC macro for trapping Ctrl-C in scripts
+ [155]Date-time parsing improvements, timezones, comparison,
arithmetic
+ [156]Pattern-matching improvements
+ FTP improvements
+ SET EXIT HANGUP { ON, OFF }
+ SET FILE EOF { CTRL-Z, LENGTH }
+ ASK[Q] /TIMEOUT
+ Bug fixes
+ New platforms
C-Kermit 8.0.200 Beta.04 (16 Nov 2001)
+ [157]New Unix man page
+ [158]New Unix installation instructions
+ SET TELOPT policies are now enforced on non-Telnet ports if
the server begins Telnet negotiations.
+ SET TERMINAL IDLE-ACTION { TELNET-NOP, TELNET-AYT }.
+ UUCP lockfile creation race condition fixed.
+ Dialout, modem signals, hangup, hardware flow control, etc,
tested extensively on many platforms, numerous problems fixed.
+ Improved hints when dialing fails.
+ SET STOP-BITS 2 can now be given without SET FLOW HARDWARE.
+ Major improvements in RFC 2217 Telnet Com-Port Control.
+ Improved ability to REDIAL a modem server port.
+ kermit -h now shows the command name in the usage usage
string.
+ kermit -h now shows ALL command-line options.
+ kermit -s blah, where blah is a symlink, now works.
+ --noperms command-line option = SET ATTRIBUTE PERMISSIONS OFF.
+ HTTP and HTTPS URLs now supported on the command line.
+ An http command-line personality is now available.
+ Initialization file streamlined to load faster, anachronisms
removed.
+ Updated NEWS, INTRO, HELP text, SHOW commands. In particular,
see SHOW COMM, HELP SET LINE, HELP WAIT.
+ Date/time arithmetic routines converted from floating-point to
integer arithmetic (internally) for greater accuracy and
portability.
+ Quoted strings containing commas no longer break macro
execution.
+ Dynamic Kermit file-transfer timeouts are now much more
aggressive.
+ New "hot keys" to turn debug.log on/off during file transfer.
+ Improved hints when file transfer fails.
+ FTP CD orientation messages are now printed.
+ -R now accepted on the FTP command line to request Recursion.
+ -m allows Active or Passive mode to be chosen on the FTP
command line.
+ -dd on the FTP command line creates a timestamped debug.log.
+ FTP command-line security options filled in.
+ Improved automatic text/binary mode switching for MGET.
+ Removed spurious error messages that sometimes occur during
MGET.
+ DIRECTORY, GREP, TYPE, HEAD, and TAIL now have a /OUTPUT:file
option.
+ TYPE /NUMBER adds line numbers.
+ CAT = TYPE /NOPAGE; MORE = TYPE /PAGE.
+ GETOK ?-help fixed.
+ \v(timestamp) (= "\v(ndate) \v(time)")
+ \v(hour) (hour of the day, 0-23)
+ \funix2dospath() converts a UNIX path (/) to a DOS one (\).
+ \fdos2unixpath() converts a DOS (Windows, OS/2) path to a UNIX
one.
+ \fkeywordval() parses name=value pair, allows macro keyword
parameters.
+ We now make every attempt to not write passwords to the
debug.log.
+ New Certificate Authority certificates file, includes the
Kermit Project at Columbia University so you can access our
IKSD securely.
+ Secure targets improved and better documented in Unix
makefile.
+ All Linux (libc and glibc) builds consolidated under "make
linux".
+ HP-UX makefile targets now have consistent names.
+ New aix50 and aix51 targets added.
C-Kermit 8.0.200 Final (12 Dec 2001)
+ Remote/local-mode confusion on some platforms introduced in
Beta.04, fixed.
+ Many of the makefile targets adjusted, new ones added.
+ New "make install" target should please most people.
+ New command: SHOW IKSD.
+ FTP over TLS.
+ Last-minute touchups to text messages, HELP text, etc.
+ Enable modem-signal reading for SCO OSR5 and Unixware 7.
+ Special superfast TRANSMIT /BINARY /NOECHO /NOWAIT mode added.
+ Fixed PBX dialing in unmarked-area-code case.
+ Improved SHOW COMMUNICATIONS tells lockfile directory, typical
dialout device name.
+ Some FTP OPEN command parsing problems fixed.
+ Some errors in date arithmetic fixed.
+ New command: SET TERMINAL AUTODOWNLOAD { ..., ERROR { STOP,
CONTINUE } }
+ New command: HELP FIREWALL.
+ SET MODEM HANGUP-METHOD DTR added as synonym for RS232-SIGNAL
+ Support for secure URL protocols added: telnets:, ftps:,
https:.
C-Kermit 8.0.201 (8 Feb 2002)
+ Installability as an [159]SSH v2 Subsystem.
+ [160]SET LOCUS command.
+ [161]L-versions of CD, DIR, DELETE, MKDIR, etc, to force local
execution.
+ [162]USER and ACCOUNT added as synonyms for FTP USER and FTP
ACCOUNT.
+ [163]SHOW VARIABLES now accepts a list of variables.
+ Rudimentary support for [164]Caller ID when receiving phone
calls.
+ Up/Down [165]Arrow-key navigation of command history buffer.
+ [166]Automatic execution of customization file if init file is
missing.
C-Kermit 8.0.206 Beta.01 (11 Oct 2002)
New commands:
o ORIENTATION lists location-related variables and their
values.
o KCD changes to special directories by their symbolic
names ("kcd ?" for a list).
o SET CD HOME path to specify home directory for CD and KCD
commands.
o CONTINUE given at top level is equivalent to END -- handy
when PROMPT'ed out of a script, to continue the script.
New switches or operands for existing commands:
o GETOK /TIMEOUT
o ASK, ASKQ, GETOK /QUIET (suppresses error message on
timeout)
o COPY /APPEND now allows concatenating multiple source
files into one dest file.
o SET TCP { HTTP-PROXY, SOCKS-SERVER } /USER, /PASSWORD.
o DIRECTORY command now accepts multiple filespecs, e.g.
"dir a b c".
SET QUIET ON now also applies to:
o SET HOST connection progress messages.
o "Press the X or E key to cancel" file-transfer message.
o REMOTE CD response.
o REMOTE LOGIN response.
Improvements and new features:
o Numerous FTP client fixes and new features, listed below.
o C-Kermit, when in remote mode at the end of a file
transfer, now prints a one-line "where" message. Control
with SET TRANSFER REPORT.
o Unix makefile "install" target now creates an UNINSTALL
script.
o Improved operation and performance on RFC 2217 Telnet
connections.
o Improved CONNECT (interactive terminal connection)
performance.
o HELP text updated for many commands.
New or fixed makefile targets:
o Solaris 9 (several variations)
o Concurrent PowerMAX
o Mac OS X 10.2
o FreeBSD 1.0
o FreeBSD 4.6, 5.0
o AIX 5.2, 5.3
Bugs fixed (general):
o Failure to run in VMS Batch fixed.
o LDIRECTORY fixed to run Kermit's built-in DIRECTORY
command rather than an external one.
o Fixed Solaris and other SVORPOSIX builds to find out
their full hostnames rather than just the "uname -n"
name.
o Fixed some problems matching strings that start with ".".
o Fixed some problems matching pattern that contain {a,b,c}
lists.
o Fixed erroneous reporting of text-mode reception as
binary when sender did not report the file size (cosmetic
only).
o Many problems with SWITCH statements fixed.
o Fixed SET OPTIONS DIRECTORY /DOTFILES to work for server
too.
o Fixed DELETE to print an error message if the file was
not found.
o Fixed SET CONTROL UNPREFIX ALL and SET PREFIXING NONE to
do the same thing.
o Fixed bugs executing macros from within the ON_EXIT
macro.
o \fday() and \fnday() fixed for dates prior to 17 Nov
1858.
o Serial speed-changing bug in Linux fixed.
o "Unbalanced braces" script parsing errors when using
\{number} fixed.
o "if defined \v(name)" fixed to behave as described in the
book.
o Fixed Problems caused by LOCAL variables whose names are
left substrings of macro names.
o The INPUT command was fixed to honor the PARITY setting.
o Fixed bug with COPY to existing file that is longer than
source file.
o REINPUT command failed to strip braces/quotes around its
target string.
o Network directory lookups didn't work for SSH
connections.
o REMOTE SET { FILE, TRANSFER } CHARACTER-SET fixed.
o Closed some holes whereby an incompletely received file
was not deleted when SET FILE INCOMPLETE is DISCARD, e.g.
when the Kermit is hung up upon.
o SET XFER CHARACTER-SET TRANSPARENT fixed to do the same
as SET XFER TRANSLATION OFF.
o SET HOST PTY (e.g. SSH) connection fixed to pass along
window-size changes.
o C-Kermit search path for TAKE files was accidentally
disabled.
FTP client bugs fixed:
o Character set translation was broken on little-endian
(e.g. PC) architectures.
o FTP PUT /SERVER-RENAME:, /RENAME-TO:, /MOVE-TO: switches
were sticky.
o Make SET TRANSFER MODE MANUAL apply to FTP.
o Make SET FILE INCOMPLETE { KEEP, DISCARD } apply to FTP.
o FTP MGET /UPDATE handled equal times incorrectly.
o FTP MGET /RECOVER fixed to ignore file dates, use only
size.
o FTP MGET /RECOVER sometimes downloaded files it didn't
need to.
o FTP downloads with TRANSFER DISPLAY BRIEF could give
misleading error messages.
o FTP MGET temp file not deleted if FTP DEBUG set to OFF
after it was ON.
o LOCUS not switched back when FTP connection is lost.
o Set incoming file date even if it was not completely
received.
o FTP MGET sent SIZE and MDTM commands even when it didn't
have to.
o FTP MGET sent SIZE and MDTM commands even when it knew
they wouldn't work.
o FTP MGET failed if no files were selected for download.
o FTP MGET a* b* c* would fail to get any c*'s if no b*'s
existed.
o Big problems canceling MGET with Ctrl-C.
o Some extraneous LOCUS dialogs squelched.
o Some inconsistencies in SET FTP FILENAMES AUTO fixed.
o Fixed file-descriptor pileup after multiple MGETs when
using mkstemp().
o Fixed "mget foo", where foo is a directory name.
FTP improvements:
o New [167]FTP protocol features added (FEAT, MLSD).
o FTP MGET /RECURSIVE now works as expected if server
supports MLSD.
o FTP MGET /DATES-DIFFER to download if local and remote
file dates differ.
o FTP DATES default changed to ON.
o FTP MPUT, MGET /EXCEPT now allows up to 64 patterns (up
from 8).
o Top-level SITE and PASSIVE commands added for
convenience.
o MGET /COLLISION:APPEND /AS-NAME:newfile *.* puts all
remote files into one local file.
o SET FTP SERVER-TIME-OFFSET for when server has wrong
timezone set.
o Allow for alternative server interpretations of [M]MPUT
/UNIQUE.
o SET FTP ANONYMOUS-PASSWORD lets you specify the default
anonymous password.
o Allow "GET /RECURSIVE path/file" to force local
subdirectory creation.
o SET FTP DISPLAY is like SET TRANSFER DISPLAY but applies
only to FTP.
o FTP { ENABLE, DISABLE } new-protocol-feature-name.
o FTP MGET /NODOTFILES.
o Debug log now records FTP commands and responses in
grep-able format.
[ [168]Top ] [ [169]Contents ] [ [170]C-Kermit ] [ [171]Kermit Home ]
1. FIXES SINCE VERSION 7.0.196 First, the changes from 7.0.196 to 7.0.197...
Source and makefile tweaks to get successful builds on platforms that were
not available in time for the 7.0 release:
* 4.2BSD
* 4.3BSD
* AIX 4.3
* AT&T 3B2 and 3B20
* BeOS 4.5
* CLIX
* Interactive UNIX System V/386 R3.2 V4.1.1
* OS-9/68000
* OSF/1 1.3.
* PS/2 AIX 1.2.1
* SCO OSR5.0.x
* SCO Xenix 2.3.4
* SINIX 5.41/Intel
* Stratus FTX
* Stratus VOS
* SunOS 4.1 with X.25
* Ultrix 4.2
* Unixware 2.0
There were no functional changes from 196 to 197.
Fixes applied after C-Kermit 7.0.197 was released:
Source code: Big flexelint and "gcc -Wall" audit and cleanup.
Configuration:
* Solaris RTS/CTS (hardware flow control) didn't work.
* BSDI RTS/CTS worked only in one direction.
* FreeBSD 4.0 with ncurses 5.0 broke interactive command parsing.
* QNX-32 build lacked -DBIGBUFOK so couldn't execute big macros.
Connections:
* SET HOST /PTY didn't work on some platforms.
* Broken SET HOST /USER:xxx /PASSWORD:yyy /ACCOUNT:zzz switches
fixed.
* Transparent printing was broken in Unix.
* ANSWER 0 (wait forever) didn't work.
* Some problems in Multitech modem command strings.
* Spurious "?Sorry, can't condition console terminal" errors.
* Disabling modem command strings by setting them to nothing broke
dialing.
* SET DIAL TIMEOUT value was usually ignored.
* SET DIAL METHOD PULSE didn't work.
* Certain modem commands, if changed, not refreshed if modem type
changed.
* SET SESSION-LOG command was missing from VMS.
* VMS session log format fixed for scripts.
* HANGUP by dropping DTR didn't work in NetBSD.
* SET FLOW /AUTO versus SET FLOW confusion fixed.
* Spurious secondary Solaris lockfile removed.
* SCO OSR5 DTR On/Off hangup.
* UUCP lockfile race condition.
Commands and scripts:
* Missing CAUTIOUS and FAST commands restored.
* Broken PTY command in late-model Linuxes fixed (API changed).
* Fixed off-by-one error in command recall when switching direction.
* Fixed recall of commands that contain '?'.
* COPY /SWAP-BYTES didn't work on some architectures.
* Various combinations of COPY switches didn't work.
* Various problems with COPY or RENAME with a directory name as
target.
* SHIFT didn't decrement \v(argc) if used within IF, ELSE, or SWITCH
block.
* SHIFT didn't affect the \%* variable.
* Divide by zero improperly handled in some \function()s.
* Problems with RETURN from right-recursive functions.
* FSEEK /LINE \%c LAST didn't work if already at end.
* Some buffer vulnerabilities and potential memory leaks were
discovered and fixed.
* \frdirectory() fixed not to follow symbolic links.
* SET EXIT WARNING OFF fixed to work when EXIT given in a script.
* Missing DELETE and MKDIR error message fixed.
* \fday() core dump for ancient dates fixed.
File transfer:
* SEND /COMMAND was broken.
* CRECEIVE was broken (but RECEIVE /COMMAND was OK).
* Quoting wildcard chars in filenames didn't work.
* Problems canceling streaming file transfers with X or Z.
* Problems shifting between streaming and windowing file transfer.
* Non-FULL file-transfer displays erroneously said STREAMING when
not.
* An active SEND-LIST prevented GET from working.
* SET SERVER GET-PATH interpretation of relative names like "." was
wrong.
* The MAIL command was broken.
* "kermit -s *" might have skipped some files.
* Transaction log entries were not made for external protocol
transfers.
* File count report fixed to show number of files actually
transferred.
* Fixed filename conversion to convert spaces to underscores.
* Made SET PREFIXING / SET CONTROL PREFIX also adjust CLEARCHANNEL.
* More "Receive window full" errors fixed.
* Broken terminal buffering after curses display in Solaris fixed.
* SET FILE INCOMPLETE DISCARD did not work in all cases.
* Packet log changed to reformat the start-of-packet character
printably.
* Dynamic timeouts could grow ridiculously large.
Character sets:
* Hebrew-7 translations missed the letter Tav.
* C1 area of CP1252 was ignored.
* SET TRANSFER CHARACTER-SET TRANSPARENT could give garbage
translations.
* TRANSLATE might not work on Little Endian architectures.
* Insufficient range checking in certain TRANSLATE operations.
The following bugs in C-Kermit 8.0.200 were fixed in 8.0.201:
* An obscure path through the code could cause the Unix version of
C-Kermit to dump core during its startup sequence. This happened to
only one person, but now it's fixed.
* When C-Kermit 8.0 is in Kermit server mode and the client says "get
blah", where blah (on the server) is a symlink rather than a real
file, the server unreasonably refused to send the linked-to file.
* When C-Kermit is an FTP client and says "get foo/bar" (i.e. a
filename that includes one or more path segments), it failed to
accept the incoming file (this happened only with GET, not MGET).
* Array references should be case insensitive but only lowercase
array letters were accepted.
* SHOW VARIABLES dumped core on \v(sexpression) and \v(svalue).
* Spurious refusals of remote directory listings if the remote
server's date was set in the past.
* In AIX, and maybe elsewhere too, Kermit's COPY command always
failed with "Source and destination are the same file" when the
destination file didn't exist.
* The VMS version of C-Kermit did not work in Batch or when SPAWN'd.
To compound the problem, it also pretty much ignored the -B and -z
command-line options, whose purpose is to work around such
problems.
* C-Kermit 8.0 could not be built on IRIX 5.x.
* The C-Kermit 8.0 build for QNX6 said it was an "(unknown version)".
Other fixes are listed in the [172]previous section.
[ [173]Top ] [ [174]Contents ] [ [175]C-Kermit ] [ [176]Kermit Home ]
2. SSH AND HTTP
2.1. SSH Connections
This section does not apply to [177]Kermit 95 2.0, which has its own
built-in SSH client, which is documented [178]SEPARATELY.
On most UNIX platforms, C-Kermit can make SSH (Secure SHell) connection
by running the external SSH command or program through its
pseudoterminal interface. The command is:
SSH text
Tells Kermit to start the external SSH client, passing the given
text to it on the command line. Normally the text is just the
hostname, but it can be anything else that is acceptable to the
ssh client. If the command succeeds, the connection is made and
Kermit automatically enters CONNECT (terminal) mode. You can use
the SSH command to make a connection to any host that has an SSH
server.
Kermit's SSH command gives you all the features of Kermit on an SSH
connection: command language, file transfer, character-set translation,
scripting, and all the rest. By default, C-Kermit invokes SSH with "-e
none", which disables the ssh escape character and makes the connection
transparent for purposes of file transfer. You can, however, change the
SSH invocation to whatever else you might need (an explicit path,
additional command-line arguments, etc) with:
SET SSH COMMAND text
Specifies the system command that Kermit's SSH command should
use to invoke the external SSH client. Use this command to
supply a specific path or alternative name, or to include
different or more command-line options.
In most cases, these connections work quite well. They can be scripted
like any other connection, and file transfer goes as fast as, or faster
than, on a regular Telnet connection. In some cases, however, the
underlying pseudoterminal driver is a limiting factor, resulting in
slow or failed file transfers. Sometimes you can work around such
problems by reducing the Kermit packet length. Note that Kermit does
not consider SSH connections to be reliable, so it does not offer to
use streaming in Kermit protocol transfers (but you can force it with
SET RELIABLE or SET STREAMING if you wish).
The SSH command is like the TELNET command: it enters CONNECT mode
automatically when the connection is made. Therefore, to script an SSH
connection, use:
set host /pty ssh -e none [ other-options ] host
if fail ...
to make the connection.
Here's a sequence that can be used to make a connection to a given host
using Telnet if the host accepts it, otherwise SSH:
if not defined \%1 exit 1 Usage: \%0 host
set quiet on
set host \%1 23 /telnet
if fail {
set host /pty ssh -l \m(user) -e none \%1
if fail exit 1 \%1: Telnet and SSH both fail
echo SSH connection to \%1 successful
} else {
echo Telnet connection to \%1 successful
}
In SSH v2, it is possible to make an SSH connection direct to a Kermit
server system if the host administrator has configured the SSH server
to allow this; [179]CLICK HERE for details.
Since Kermit uses external ssh client software, and since there are
different ssh clients (and different releases of each one), the exact
command to be used to make an SSH/Kermit connection can vary. Here is
the command for the OpenSSH 3.0.2p1 client:
set host /pipe ssh -e none [ -l username ] -T -s hostname kermit
Example:
set host /pipe ssh -e none -l olga -T -s hq.xyzcorp.com kermit
The SSH client might or might not prompt you for a password or other
information before it makes the connection; this depends on your SSH
configuration (your public and private keys, your authorized hosts
file, etc). Here's a brief synopsis of the OpenSSH client command
syntax ("man ssh" for details):
-e none
This tells the SSH client to use no escape character. Since we
will be transferring files across the connection, we don't want
the connection to suddenly block because some character in the
data.
-l username
This is the username on the remote host. You can omit the -l
option and its argument if your local and remote usernames are
the same. If they are different, you must supply the remote
username.
-T
This tells the SSH client to tell the SSH server not to allocate
a pseudoterminal. We are not making a terminal connection, we
don't need a terminal, and in fact if a terminal were allocated
on the remote end, the connection would not work.
-s ... kermit
This tells the SSH client to tell the SSH server to start the
specified subsystem ("kermit") once the connection is made. The
subsystem name comes after the hostname.
hostname
The IP host name or address of the desired host.
You might want to include other or additional ssh command-line options;
"man ssh" explains what they are. Here are some examples for the
OpenSSH 3.0.2p1 client:
-oClearAllForwardings yes
-oForwardAgent no
-oForwardX11 no
-oFallbackToRsh no
These ensure that a secure connection is used and that the
connection used for file transfer is not also used for
forwarding other things that might be specified in the
ssh_config file.
-oProtocol 2
(i.e. SSH v2) Ensures that the negotiated protocol supports
subsystems.
Once you have an SSH connection to a Kermit server, it's just like any
other connection to a Kermit server (and very similar to a connection
to an FTP server). You give the client file transfer and management
commands for the server, and the server executes them. Of course you
can also give the client any other commands you wish.
[ [180]SSH Kermit Server Subsystem ] [ [181]Kermit 95 Built-in SSH
Client ]
2.2. HTTP Connections
Hypertext Transfer Protocol, or HTTP, is the application protocol of
the World Wide Web (WWW), used between Web browsers (clients) and Web
servers. It allows a client to get files from websites, upload files to
websites, delete files from websites, get information about website
directories and files, and interact with server-side CGI scripts.
C-Kermit includes an HTTP client capable of both clear-text and secure
HTTP connections, that can do all these tasks and can be automated
through the Kermit scripting language.
Although C-Kermit 7.0 could make HTTP connections to Web servers, it
could do so only when no other connection was open, and the procedure
was somewhat awkward. C-Kermit 8.0 improves matters by:
* Allowing an HTTP connection to be open at the same time as a
regular SET LINE or SET HOST connection, and also at the same time
as an FTP connection ([182]Section 3);
* Upgrading the HTTP protocol level from 1.0 to 1.1, thus allowing
for persistent connections, in which a series of commands can be
sent on the same connection, rather than only one as in HTTP 1.0
(and C-Kermit 7.0);
* Providing for "one-shot" URL-driven HTTP operations such as GET or
PUT.
* Providing a distinct HTTP command-line personality.
Persistent HTTP connections are managed with the following commands:
HTTP [ switches ] OPEN [ security-options ] host-or-url [ port ]
Opens a persistent connection to the specified host (IP host
name or address) on the specified port. If any switches
(options, listed in the next section) are included, their values
are saved and used for all subsequent HTTP action commands on
the same connection. If no port is specified, HTTP (80) is used.
A Uniform Resource Locator (URL, [183]RFC 1738) can be given
instead of a hostname (or address) and port (but the URL can not
include a directory/file path). The security options are
explained [184]below. The HTTP OPEN command replaces the
C-Kermit 7.0 SET HOST hostname HTTP command, which no longer
works with HTTP GET and related commands.
HTTP CLOSE
Closes any open HTTP connection and clears any saved switch
values.
A URL starts with a protocol name, which must be http or https in this
case; optionally includes a username and password; and must contain a
host name or address:
protocol://[user[.password]]@host[:port][URI]
HTTP is Hypertext Transfer Protocol. HTTPS is the secure (SSL/TLS)
version of HTTP. The TCP service port is derived from the protocol
prefix (so normally the ":port" field is omitted). Thus the URL
protocol name specifies a default TCP service port and the URL user and
password fields can take the place of the /USER and /PASSWORD switches
([185]Section 2.2.1). The optional URI is a "compact string of
characters for identifying an abstract or physical resource" ([186]RFC
2396), such as a file. It must begin with a slash (/); if the URI is
omitted, "/" is supplied. Examples:
http open http://www.columbia.edu/
Equivalent to http open www.columbia.edu or http open
www.columbia.edu http.
http open https://olga.secret@www1.xyzcorp.com/
Equivalent to http /user:olga /pass:secret open www1.xyzcorp.com
https.
Persistence is accomplished unilaterally by C-Kermit 8.0. An HTTP 1.0
server closes the connection after each action. Although HTTP 1.1
allows multiple actions on the same connection, an HTTP 1.1 server
tends to close the connection if it is idle for more than a few
seconds, to defend itself against denial-of-service attacks. But when
you use Kermit's HTTP OPEN command to create a connection, Kermit
reopens it automatically (if necessary) for each HTTP action until you
close it with HTTP CLOSE, regardless of the server's HTTP protocol
version, or how many times it closes the connection.
Firewalls can be negotiated through proxies with the following
commands:
SET TCP HTTP-PROXY [ host[:port] ]
If a host (by hostname or IP address) is specified, Kermit uses
it as a proxy server when attempting outgoing TCP connections --
not only HTTP connections, but all TCP/IP connections, Telnet
and FTP included. This allows Kermit to adapt to the HTTP
firewall penetration method (as opposed to other methods such as
SOCKS4). If no hostname or ip-address is specified, any
previously specified Proxy server is removed. If no port number
is specified, the "http" service is used. This command must be
given before the HTTP OPEN command if a proxy is to be used or
canceled.
HTTP [ switches ] CONNECT host[:port]
Instructs the HTTP server to act as a proxy, establishing a
connection to the specified host (IP hostname or address) on the
given port (80 = HTTP by default) and to redirect all data
transmitted between Kermit and itself to the given host for the
life of the connection. This command is to be used only for
debugging HTTP proxy connections. If a proxy connection is
required, instruct Kermit to use the proxy with the SET TCP
HTTP-PROXY command.
2.2.1. HTTP Command Switches
HTTP switches, like all other switches, are optional. When HTTP
switches are included with the HTTP OPEN command, they apply
automatically to this and all subsequent HTTP actions (GET, PUT, ...)
on the same connection until an HTTP CLOSE command is given. So if you
include switches (or the equivalent URL fields, such as user and
password) in the HTTP OPEN command, you can omit them from subsequent
commands on the same connection. If the connection has closed since
your last command, it is automatically reopened with the same options.
If you include switches with an HTTP action command (such as GET or
PUT), they apply only to that command.
/USER:name
To be used in case a page requires a username for access. The
username is sent with page requests. If it is given with the
OPEN command it is saved until needed. If a username is included
in a URL, it overrides the username given in the switch.
CAUTION: Username and password (and all other information,
including credit card numbers and other material that you might
prefer to protect from public view) are sent across the network
in clear text on regular HTTP connections, but authentication is
performed securely on HTTPS connections.
/PASSWORD:text
To be used in case a web page requires a password for access.
The password is sent with page requests. If it is given with the
OPEN command it is saved until needed. If a password is given in
a URL, it overrides the one given here. CAUTION: (same as for
/USER:).
/AGENT:user-agent
Identifies the client to the server. Overrides the default agent
string, which is "C-Kermit" (for C-Kermit) or "Kermit-95" (for
Kermit 95).
/ARRAY:array-designator
Tells Kermit to store the response headers in the given array,
one line per element. The array need not be declared in advance.
Example: /array:&a.
/TOSCREEN
Tells Kermit to display any response text on the screen. It
applies independently of the output file specification; thus it
is possible to have the server response go to the screen, a
file, both, or neither.
/HEADER:header-item(s)
Used for specifying any optional headers to be sent with HTTP
requests.
/HEADER:tag:value
To send more than one header, use braces for grouping:
/HEADER:{{tag:value}{tag:value}...}
For a list of valid tags and value formats see [187]RFC 2616,
"Hypertext Transfer Protocol -- HTTP/1.1". A maximum of eight
headers may be specified.
2.2.2. HTTP Action Commands
HTTP actions can occur within a persistent connection, or they can be
self-contained ("connectionless"). A persistent HTTP connection begins
with an HTTP OPEN command, followed by zero or more HTTP action
commands, and is terminated with an HTTP CLOSE command:
http open www.columbia.edu
if failure stop 1 HTTP OPEN failed: \v(http_message)
http get kermit/index.html
if failure stop 1 HTTP GET failed: \v(http_message)
(more actions possible here...)
http close
A self-contained HTTP action occurs when a URL is given instead of a
remote file name to an HTTP action command. In this case, Kermit makes
the HTTP connection, takes the action, and then closes the connection.
If an HTTP connection was already open, it is closed silently and
automatically.
http get http://www.columbia.edu/kermit/index.html
Kermit's HTTP action commands are as follows. Switches may be included
with any of these to override switch (or default) values given in the
HTTP OPEN command.
HTTP [ switches ] GET remote-filename [ local-filename ]
Retrieves the named file from the server specified in the most
recent HTTP OPEN command for which a corresponding HTTP CLOSE
command has not been given. The filename may not include
wildcards (HTTP protocol does not support them). If no HTTP OPEN
command is in effect, this form of the HTTP GET command fails.
The default local filename is the same as the remote name, but
with any pathname stripped. For example, the command http get
kermit/index.html stores the file in the current local directory
as index.html. If the /HEADERS: switch is included, information
about the file is also stored in the specified array (explained
in [188]Section 2.2.3). All files are transferred in binary
mode. HTTP does not provide for record-format or character-set
conversion.
HTTP [ switches ] GET url [ local-filename ]
When HTTP GET is given a URL rather than a filename, Kermit
opens a connection to the designated server (closing any
previously open HTTP connection), gets the file, and then closes
the connection. If the URL does not include a filename,
index.html is supplied. This is the self-contained one-step
"connectionless" method for getting a file from a Web server.
The data is not interpreted; HTTP GET is like "lynx -source"
rather than "lynx -dump".
In the remaining HTTP action commands, the distinction between a remote
filename and a URL are the same as in the HTTP GET command.
HTTP [ switches ] HEAD remote-filename-or-url [ local-filename ]
Like GET except without actually getting the file; instead it
retrieves only the headers. If the /ARRAY: or /TOSCREEN switch
is included, there is no default local output filename but you
can still specify one. If neither of these switches is included,
the default local filename is the same as the remote filename,
but with any path stripped and with ".head" appended. The HEAD
command can be used in a script with the /ARRAY: switch to
retrieve information about the requested resource to determine
whether the resource should actually be retrieved with a
subsequent GET request.
HTTP [ switches ] INDEX remote-directory-or-url [ local-filename ]
Asks the server to send a listing of the files in the given
server directory. This command is not supported by most Web
servers. Even when it is supported, there is no standard format
for the listing.
HTTP [ switches ] POST [ /MIME-TYPE:type ] source-file
remote-path-or-url [ result-file ]
Sends data to a process running on the remote host; the result
is usually an HTML file but could be anything. The data to be
posted must be read from a local file (the source-file). If a
result file is specified, Kermit stores the server's response in
it.
HTTP [ switches ] PUT [ MIME-TYPE:type ] local-file [
remote-file-or-url [ result-file ] ]
Uploads a local file to the server. Only the name of a single
file can be given; wildcards (and group transfers) are not
supported by HTTP protocol. If no remote filename is given, the
file is sent with the same name as the local file, but with any
pathname stripped.
HTTP [ switches ] DELETE remote-file-or-url [ local-result-file ]
Asks the server to delete the specified single file. If a result
file is specified, it will contain any response data returned by
the server.
Note the limitations of HTTP protocol compared to (say) FTP or Kermit.
There is no command for changing directories, no standard way to get
file or directory lists, no way to transfer file groups by using
wildcard notation, etc, and therefore no good way to (say) fetch all
pages, descend through subdirectories, perform automatic updates, etc.
There is no assurance a connection will stay open and, as noted, there
is no provision for data conversion between unlike platforms. The
data's MIME headers can be used for postprocessing.
2.2.3. HTTP Headers
Each HTTP request and response contains a set of name/value pairs
called headers. HTTP headers are specified in [189]RFC 2616. For
example, an HTTP GET request for /index.html on www.columbia.edu
contains the following headers:
GET /index.html HTTP/1.1
Host: www.columbia.edu:80
User-agent: C-Kermit 8.0
Authorization: Basic base64-encoded-username-password
These might be followed by any others specified with a /HEADERS:
switch:
Accept: image/gif, image/x-xbitmap, image/jpeg, *.*
Accept-Encoding: gzip
Accept-Language: en
Accept-Charset: iso-8859-1,utf-8
Cookie: cookie-data
The server sends back a short report about the file prior to sending
the file contents. Example:
HTTP/1.1 200 OK
Date: Fri, 24 Aug 2001 21:09:39 GMT
Server: Apache/1.3.4 (Unix)
Last-Modified: Mon, 06 Aug 2001 21:16:13 GMT
ETag: "1fa137-10d7-3b6f091d"
Accept-Ranges: bytes
Content-Length: 4311
Content-Type: text/html
If you want to have this information available to a Kermit script you
can use the /ARRAY switch to have Kermit put it in array, one line per
array element. Example:
set exit warning off
http open www.columbia.edu
if fail exit 1 Can't reach server
http /array:&a get /index.html
if fail exit 1 Can't get file
echo Header lines: \fdim(&a)
for \%i 1 \fdim(&a) 1 {
echo \%i. \&a[\%i]
}
Note that the "Date:" item is the current date and time; the
"Last-Modified:" item is the file's modification date and time. An
example showing how to use this information is presented in
[190]Section 8.13.7.
2.2.4. Secure HTTP Connections
SSL/TLS (Secure Sockets Layer / Transport Layer Security) is the
protocol used to secure HTTP, SMTP, and other Internet applications.
See the [191]C-Kermit Reference Section 5.4 for an introduction to
SSL/TLS. To make a secure HTTP connection, you need:
1. A secure client (a version of C-Kermit or Kermit 95 with SSL/TLS
security built in). Type "check ssl" at the Kermit prompt to make
sure you have it.
2. A secure server to connect to.
3. The CA Root Certificate used to authenticate the server to the
client. (see [192]Section 15 of the security reference for an
introduction to certificates).
And you must make a connection to the secure HTTP port: service name
HTTPS, port number 443 (as opposed to service HTTP, port 80). You can
also make secure connections to other ports by including the /TLS or
/SSL switch with the HTTP OPEN command, if the host supports SSL/TLS on
the given port:
The quality of the SSL/TLS connection depends on the cipher suite.
There are several possibilities:
Anonymous cipher suite:
If an anonymous cipher suite is negotiated, the connection is
encrypted but there is no authentication. This connection is
subject to a Man-In-The-Middle (MITM) attack.
X.509 certificate on the server:
When you connect to certain secure servers, an X.509 certificate
is returned. This certificate is issued to a special hostname,
something like www1.xyzcorp.com or wwws.xyzcorp.com (rather than
the normal www.xyzcorp.com). It is signed by the host's
Certificate Authority (CA). If the host certificate is
configured on the client, it can be used to verify the
certificate received from the server. If the certificate it
verified as authentic, a check is made to ensure it has not
expired and it was issued to the host you were attempting to
connect to. If you had asked to connect to (say) www.xyzcorp.com
but were given a certificate for www1.xyzcorp.com, you would be
prompted for permission to continue.
If the verification succeeded, the connection would be encrypted
with one-way (server-to-client) authentication. This connection
is not subject to a MITM attack.
If a username and password are transmitted over this connection,
they are not subject to interception. However, the standard
risks associated with passing the password to the host for
verification apply; for example, if the host has been
compromised, the password will be compromised.
X.509 client certificate:
If a connection has been established with an X.509 server
certificate, the server can ask the client to send a certificate
of its own. This certificate must be verified against a CA Root
certificate. The certificate itself (or subject info from the
certificate) is used to determine the authorization for the
client, and if successful, the username and password need not be
sent to the server.
Kerberos 5:
Instead of using X.509 certificates, Kerberos 5 can be used to
perform the authentication and key exchange. In this situation,
there is mutual authentication between the client and server.
The Kerberos 5 principal is used by the server to look up the
appropriate authorization data. There is no need to send
username and password.
An HTTP connection is made with the HTTP OPEN command:
HTTP [ switches ] OPEN [ { /SSL, /TLS } ] host [ port ]
If /SSL or /TLS switches are included (these are synonyms), or
if the service is HTTPS or the port is 443, a secure connection
is attempted using the current authentication settings; see HELP
SET AUTHENTICATION for details ([193]Section 6.2 of the security
reference). If the no /SSL or /TLS switch is included but the
port is 443 or the service is HTTPS, a secure connection is
attempted. If an /SSL or /TLS switch is included but a port is
not specified, an SSL/TLS connection is attempted on the default
port (80).
Certificates are covered in the separate [194]Kermit Security Reference
for C-Kermit 8.0. You should let Kermit know to verify certificates
with the SET AUTHENTICATION TLS command. For example:
SET AUTHENTICATION TLS CRL-DIR directory
Specifies a directory that contains certificate revocation files
where each file is named by the hash of the certificate that has
been revoked.
SET AUTHENTICATION TLS CRL-FILE filename
Specifies a file that contains a list of certificate
revocations.
SET AUTHENTICATION TLS VERIFY-DIR directory
Specifies a directory that contains root CA certificate files
used to verify the certificate chains presented by the peer.
Each file is named by a hash of the certificate.
SET AUTHENTICATION TLS VERIFY-FILE filename
Specifies a file that contains root CA certificates to be used
for verifying certificate chains.
SET AUTHENTICATION TLS VERIFY OFF
Tells Kermit not to require a certificate and accept any
certificate that is presented regardless of whether it is valid.
There are many other options; see the security document for details.
Now suppose you need need to fetch the file denoted by the following
URL:
https://myuserid:mypassword@wwws.xyzcorp.com/clients/info/secret.html
Once you have set up the handling of certificates as desired, you can
use the following Kermit commands:
http /user:myuserid /password:mypassword open www1.xyzcorp.com https
if success {
http get /clients/info/secret.html
http close
}
As another example, let's say that you have a web form you need to
populate with three fields: red,white and blue.
<FORM ACTION="http://www.xyzcorp.com/cgi-bin/form.cgi" METHOD="POST">
<INPUT NAME="Red">
<INPUT NAME="White">
<INPUT NAME="Blue">
</FORM>
You can handle this with the HTTP POST command. The data to be posted
is stored in the local file data.txt.
Red=seven stripes&White=six stripes&Blue=fifty stars
and the response from the server will be stored into response.txt.
http open www.xyzcorp.com http
if success {
http /array:c post data.txt /cgi-bin/form.cgi response.txt
http close
}
In this scenario, the Common Gateway Interface (CGI) sends a response
whether it succeeds or fails in a script-dependent manner. The script
can either report success and enclose the response data; or it might
send a 302 Found error which indicates that the "Location:" header
should be used to determine the URL at which the data can be found.
2.2.5. HTTP Variables
\v(http_code)
The HTTP protocol code number of the most recent server reply,
e.g. 404 for "not found".
\v(http_connected)
1 when an HTTP connection is open, 0 when there is no HTTP
connection.
\v(http_host)
If an HTTP connection is open, the hostname:port, e.g.
www.columbia.edu:80; otherwise, empty.
\v(http_message)
Server error message, if any, from most recent HTTP command.
\v(http_security)
A list of the security parameters and values for the current
connection, if any. Empty if the connection is not to a secure
server, or there is no connection.
To display all the HTTP variables at once, type SHOW VAR HTTP:
C-Kermit> http open www.columbia.edu
C-Kermit> http get lkjlkjlkjlkj
C-Kermit> sho var http
\v(http_code) = 404
\v(http_connected) = 1
\v(http_host) = www.columbia.edu:80
\v(http_message) = Not Found
\v(http_security) = NULL
C-Kermit>
2.2.6. The HTTP Command-Line Personality
If you invoke C-Kermit with the name "http" or "https", you can use a
special set of HTTP-specific command-line options. You can do this by
creating a symbolic linke "http" or "https" to the C-Kermit 8.0
executable, or by having a separate copy of it called "http" or
"https". Here's the usage message ("http -h"):
Usage: ./http host [ options... ]
-h This message.
-d Debug to debug.log.
-S Stay (issue command prompt when done).
-Y Do not execute Kermit initialization file.
-q Quiet (suppress most messages).
-u name Username.
-P password Password.
-g pathname Get remote pathname.
-p pathname Put remote pathname.
-H pathname Head remote pathname.
-l pathname Local path for -g, -p, and -H.
-z opt[=value] Security options...
cert=file Client certificate file
certsok Accept all certificates
key=file Client private key file
secure Use SSL
verify=n 0 = none, 1 = peer , 2 = certificate required
The "host" argument is the name of a Web host, e.g. www.columbia.edu.
The action options are -p, -g, and -H. If you give an action option,
Kermit does the action and then exits. If you give a host without an
action option, Kermit makes an HTTP connection to the host and then
gives you the C-Kermit prompt. Here's a simple example that fetches a
publicly readable Web page:
http www.columbia.edu -g kermit/index.html
If you need to access a website for which a username and password are
required, you can supply them on the command line with -u and -P. If
you include a username but omit the password, Kermit prompts you for
it:
http www.columbia.edu -u olga -p kermit/index.html -l index.html
Password:
Note that when PUT'ing files to websites, you have to supply both the
-p (remote pathname) and -l (local path) options.
If your version of Kermit is built with SSL/TLS security, you can also
use the -z option to make secure HTTP (https) connections.
Finally, as noted in [195]Section 16, you can also give a URL instead
of a host name and options.
[ [196]Top ] [ [197]Contents ] [ [198]C-Kermit Home ] [ [199]Kermit
Home ]
3. KERMIT'S BUILT-IN FTP CLIENT
3.1. [200]Making and Managing FTP Connections
3.2. [201]Making Secure FTP Connections
3.3. [202]Setting FTP Preferences
3.4. [203]Managing Directories and Files
3.5. [204]Uploading Files With FTP
3.6. [205]Downloading Files With FTP
3.7. [206]Translating Character Sets
3.8. [207]FTP Command Shortcuts
3.9. [208]Dual Sessions
3.10. [209]Automating FTP Sessions
3.11. [210]Advanced FTP Protocol Features
Earlier versions of C-Kermit and K95 included an FTP command, but it
simply invoked an external FTP client. Now, by popular demand, Kermit
includes its own built-in FTP client that offers the following
advantages over traditional FTP clients (and its previous interface to
them):
* Any of Kermit's built-in [211]security methods can be used to
establish and conduct secure FTP sessions with [212]FTP servers
that support these methods. (Security modules can be subject to
export restrictions.)
* Kermit's FTP client uses "passive mode" by default to avoid
blockage by firewalls and network address translators. Of course
active mode can be chosen too when needed.
* [213]Character sets can be translated as part of the transfer
process even when the FTP server does not support character-set
translation, including to/from the new Internet standard
international character set, [214]Unicode UTF-8. This includes both
the file's name and (for text files only) its contents.
* All of C-Kermit's [215]file-selection mechanisms are available:
size, date, name patterns and lists, exception lists, etc.
* [216]Atomic file movement capabilities are provided (delete, move,
or rename files automatically after successful transfer).
* The correct file type, "ascii" (i.e. text) or binary, is chosen
automatically for each file (explained in [217]Section 4), and any
mixture of text and binary files can be sent in a single operation,
even across platforms.
* Update mode ("don't bother transferring files that didn't change
since last time") and recovery (resumption of an interrupted
transfer from the point of failure) are available in both
directions.
* When uploading files from UNIX to UNIX, the file's permissions can
be preserved if desired.
* Recursive directory-tree PUTs are supported between any two
platforms that have tree-structured file systems. Recursive GETs
are supported between like platforms if the server cooperates and
between like or unlike platforms if the server supports MLSD
([218]Section 3.11).
* When receiving files, all of Kermit's file collision actions are
available: backup, update, refuse, rename, etc.
* Multi-file transfers can be interrupted on a per-file basis,
automatically skipping to the next file.
* FTP sessions are [219]fully scriptable.
* An entire FTP session (connect, login, CD, upload or download,
logout) can be specified on the command line without using a
script.
* All of Kermit's logging options and formats are available to keep
an accurate and complete record of each connection and file
transfer, and to aid in troubleshooting.
* All of Kermit's file-transfer display options are available
(fullscreen, brief, CRT, serial, none).
And best of all:
* Kermit doesn't give you those annoying per-file prompts every time
you start a multi-file transfer without remembering to give a
"prompt" command first :-).
[ [220]Top ] [ [221]FTP Top ] [ [222]FTP Client Overview ] [ [223]FTP
Script Tutorial ] [ [224]C-Kermit Home ] [ [225]Kermit Home ]
3.1. Making and Managing FTP Connections
Each copy of Kermit can have one FTP connection open at a time. FTP
connections are independent of regular terminal connections; a terminal
connection (serial or network via SET LINE, DIAL, SET HOST, TELNET,
etc) may be, but need not be, open at the same time as an FTP
connection, and terminal connections can also be closed, and new
connections opened, without interfering with the FTP connection (and
vice versa). Thus, for example, Kermit can have an FTP connection and a
TELNET connection open to the same host simultaneously, using the
TELNET connection (e.g.) to send mail or take other desired actions as
various FTP actions complete. Of course, each copy of Kermit can do
only one thing at a time, so it can't (for example) transfer a file
with FTP and another file with Kermit protocol simultaneously.
A Kermit FTP session can be established by [226]command-line options,
by [227]URL, or by [228]interactive commands.
3.1.1. Kermit Command-Line Options for FTP
The new command-line option '-9' (sorry, we're out of letters) can be
used when starting C-Kermit, telling it to make an FTP connection:
kermit -9 hostname
or if a non-default FTP port is needed:
kermit -9 hostname:port
You can also specify the username on the command line with the -M ("My
User ID") option that was already there for other connection types:
kermit -9 hostname -M olga
If you specify the username on the command line, Kermit uses it when
making the connection and does not prompt you for it (but it does
prompt you for the password if one is required).
Once the connection is made, you get the regular Kermit prompt, and can
give interactive commands such as the ones described below. When you
give a BYE command, Kermit closes the session and exits, just as a
regular FTP client would do. If you don't want Kermit to exit when you
give a BYE command, include the -S ("Stay") option on the command line.
Other Kermit command-line options that are not specific to non-FTP
connections should affect the FTP session in the expected ways; for
example, -i and -T force binary and text mode transfers, respectively.
File transfers can not be initiated on the "kermit -9" command line;
for that you need to use Kermit's FTP personality (next section) or you
can use URLs ([229]Section 3.1.3).
3.1.2. The FTP Command-Line Personality
If you want to replace your regular FTP client with C-Kermit, you can
make a link called "ftp" to the C-Kermit binary (or you can store a
copy of the C-Kermit binary under the name "ftp"). When C-Kermit is
invoked with a program name of "ftp" (or "FTP", case doesn't matter),
it assumes the command-line personality of the regular FTP client:
ftp [ options ] hostname [ port ]
In this case the options are like those of a regular FTP client:
-d Debug: enables debug messages and creates a debug.log file.
-n No autologin: Kermit should not send your user ID automatically.
-t Packet trace: accepted but is treated the same as -d.
-v Verbose: accepted but ignored (operation is verbose by default).
-i Not interactive: accepted but ignored.
and the hostname can also be a URL (explained in [230]Section 3.1.3).
To specify a non-default TCP port for the FTP server, include the port
number or name after the hostname.
There are also some bonus options that allow you to execute an entire
FTP session from the shell command line, as long as you don't include
the -n option. These are not available with regular FTP clients, and at
least one of these options (-g) conflicts with UNIX ftp (where -g means
"no globbing", which does not apply to Kermit), and some of them (like
the options above) also conflict with regular Kermit command-line
options:
-m mode = "passive" (default) or "active"
-Y Don't execute the Kermit initialization file [1]
-q Quiet, suppresses all but error messages [1]
-S Stay, don't exit automatically [1]
-A Autologin anonymously [2]
-u name Username for autologin [2] (synonym: -M [1])
-P password Password for autologin (see cautions below) [2]
-D directory cd after autologin [2]
-b Binary mode [2]
-a Text ("ascii") mode [2] (synonym: -T [1])
-R Recursive (works with -p) [4]
-p files Files to put (upload) after autologin [2] (synonym: -s [1])
-g files Files to get (download) after autologin [3]
[1] Same as Kermit, not available in regular FTP clients.
[2] Conflicts with Kermit, not available in regular FTP clients.
[3] Same as Kermit, conflicts with regular FTP clients.
[4] Conflicts with Kermit, available in some FTP clients.
Fancier options such as restart, character-set translation, filename
collision selection, automatic move/rename/delete, etc, are not
available from the command line; for these you can use the commands
described in the following sections. The -R option might also work with
-g (GET) but that depends on the server.
The following security options are also available, explained in
[231]Section 3.2:
-k realm Kerberos 4 realm [4]
-f Kerberos 5 credentials forwarding [4]
-x autoencryption mode [4]
-c cipher SRP cipher type [4]
-H hash SRP encryption hash [4]
-z option Security options [4]
If you include -A or specify a name of "anonymous" or "ftp", you are
logged in anonymously and, in the absence of -P, Kermit automatically
supplies a password of "user@host", where "user" is your local user ID,
and "host" is the hostname of the computer where Kermit is running. If
you do not include -p or -g, Kermit enters command mode so you can type
commands or execute them from a script.
If you include -p or -g, Kermit attempts to transfer the specified
files and then exits automatically at the end of the transfer unless
you also included -S (Stay). It uses the "brief" file transfer display
(one line per file) unless you include the -q option to suppress it.
When uploading files with -p, Kermit switches automatically between
text and binary mode for each file.
When downloading, you can either specify a particular mode (text or
binary) to be used for all the files, or you can let Kermit select the
type for each file automatically, based on its name (see [232]Sections
3.5 and [233]3.6 for greater detail). In UNIX be sure to quote any
wildcard characters to prevent the shell from expanding them, as shown
in the examples just below. Filename collisions are handled according
Kermit's FILE COLLISION setting (if specified in your Kermit
customization file; otherwise the default, which is BACKUP).
It should go without saying that the -P option should be used with
caution. In addition to the well-known risks of transmitting plaintext
passwords over the Internet, in this case the password also echos to
the screen if you type it, and can be seen in ps and w listings that
show the user's currently active command and command-line arguments.
Thus command-line FTP sessions are most appropriate for secure or
anonymous connections (those that do not require passwords).
Here's an example in which you download the latest C-Kermit "tarball"
from the Columbia University FTP archive:
ftp -A kermit.columbia.edu -bg kermit/archives/ckermit.tar.gz
This assumes that "ftp" is a symbolic link to C-Kermit. It logs you in
anonymously and gets the ckermit.tar.gz file in binary mode from the
kermit/archives directory.
Here's a slightly more ambitious example that illustrates CD'ing to the
desired server directory to get a group of files in text mode (in this
case the C-Kermit source files):
ftp -A kermit.columbia.edu -D kermit/f -ag "ck[cuw]*.[cwh]" makefile
In this case we CD to the kermit/f directory so we don't have to
include it in each file specification, and we quote the ck[cuw]*.[cwh]
specification so the shell doesn't expand it, since we have to pass it
as-is to the server. Note also that the quotes don't go around the
entire file list; only around each file specification that needs to be
quoted.
Here's one more example, that uploads a debug log file in binary mode
to the Kermit incoming directory (as we might ask you to do when
following up on a problem report):
ftp -A kermit.columbia.edu -D kermit/incoming -bp debug.log
In this case the -D option is required to tell the server where to put
the incoming file.
Unless the -Y option is included, your Kermit initialization file
(.mykermrc in UNIX, K95.INI in Windows) is executed before the command
line options, so you can set any FTP-related preferences there, as
described in the subsequent sections.
3.1.3. The FTP URL Interpreter
If Kermit is invoked with either its regular personality (as "kermit")
or its FTP personality (as "ftp"), you can also give a URL
(Universal Resource Locator) instead of a hostname and options,
with or without a username and password:
ftp ftp://user:password@host/path
ftp ftp://user@host/path
ftp ftp://@host/path (or ftp://:@host/path)
ftp ftp://host/path
kermit ftp://host/path
If the FTP personality is used, the service must be "ftp". In all
cases, a hostname or address must be included. If a user is included
but no password, you are prompted for the password. If a path
(filename) is included:
* If "@" is included without a user, Kermit prompts for the username
and password.
* If no user and no "@" are included, "anonymous" is used.
* GET is assumed.
If no path (and no action options) are included, an interactive FTP
session is started, as in this example:
ftp ftp://kermit.columbia.edu
If a path is included, but a username is not included, "anonymous" is
used and an appropriate user@host password is supplied automatically.
If authentication is successful, Kermit attempts to GET the file
indicated by the path or, if the path is the name of a directory, it
asks the server for a directory listing. In both cases, Kermit
disconnects from the server and exits after the operation is complete
(unless you have included the -S option on the command line).
Here's an example that gets a listing of the Kermit directory at the
Kermit ftp site:
ftp ftp://kermit.columbia.edu/kermit/
This example gets the top-level READ.ME file from the same directory:
ftp ftp://kermit.columbia.edu/kermit/READ.ME
Here's the same example, but requesting a text-mode transfer:
ftp -T ftp://kermit.columbia.edu/kermit/READ.ME
This illustrates that you can mix command-line options and URLs
if you desire.
Here's an example that logs in as a (fictitious) real user to get a
file:
ftp ftp://olga@ftp.xyzcorp.com/resume.txt
The password is not included, so Kermit prompts for it.
This scheme allows Kermit to be used as the FTP helper of other
applications, such as Web browsers, with all its advantages over other
FTP clients (especially the ones that are built in to most Web
browsers), e.g. that it can be given wildcards, and it can pick text
and binary mode automatically for each file.
HINT: suppose somebody sends you an FTP URL in email, or you see it in
some text. If your terminal screen supports copy/paste, copy the url,
and then at the shell prompt type "kermit", a space, and then paste the
URL, e.g.:
$ kermit ftp://alpha.greenie.net/pub/mgetty/source/1.1/mgetty1.1.27-O
"$ is the shell prompt; the part you type is underlined, the rest is
pasted in. Kermit does the rest.
3.1.4. Interactive FTP Session Establishment
As you read this and the following sections, bear in mind that any
command that can be given at the prompt can also be used in a script
program. Kermit's script programming language is the same as its
interactive command language. [234]CLICK HERE if you would like to
learn a bit more about script writing.
An FTP session is established with the FTP OPEN command:
FTP [ OPEN ] [ { /SSL, /TLS } ] hostname [ switches ] [ port ]
Opens an FTP connection to the given host on the given port and,
if FTP AUTOLOGIN is ON, also logs you in to the server,
prompting for username and password if necessary. If no port is
specified, the regular FTP protocol port (21) is used. The OPEN
keyword is optional (unless the hostname conflicts with one of
the FTP command keywords, which you can list by typing "ftp ?").
The hostname can be an IP host name, numeric IP address, or if you have
a network directory active (SET NETWORK DIRECTORY; see Chapter 6 of
[235]Using C-Kermit), an entry name in the directory. In the latter
case, if the given hostname matches exactly one entry, the associated
name or address is used; if it matches more than one, Kermit cycles
through them until one is found that can be opened; if it matches none,
then the hostname is used as-is. If a directory is active but you want
to bypass directory lookup, include an "=" sign at the beginning of the
hostname, and/or use a numeric IP address.
When an FTP connection is opened, the default file-transfer mode is set
to binary if the client and server platforms are alike (e.g. both of
them are some kind of UNIX), and to text ("ascii") if they are not
alike. This has no particular effect for uploading since Kermit
automatically switches between text and binary mode for each file, but
might be important for downloading. The connection is also set to
Stream mode and File structure. Record- or page-oriented file transfers
are not supported by C-Kermit's FTP client.
The optional FTP OPEN switches are:
/ANONYMOUS
Logs you in anonymously, automatically supplying username
"anonymous" and user@host as the password, based on your local
user and host names.
/NOLOGIN
Overrides SET FTP AUTOLOGIN ON for this connection only.
/USER:name
Uses the given username to log you in, thus avoiding the Name:
prompt.
Overrides SET FTP AUTOLOGIN OFF for this connection only.
/PASSWORD:text
Uses the given text as your password, thus avoiding the
Password: prompt. This switch is not recommended for use in
script files, which would be a security risk.
/ACCOUNT:text
Uses the given text as your account (or secondary password,
depending on the requirements of the server; most servers do not
require or accept an account name). If an account is not
supplied, you are not prompted for one.
/PASSIVE
Opens the connection in passive mode. Passive mode is the
default in Kermit's FTP client, unlike in most others, since it
works better through firewalls. The /PASSIVE and /ACTIVE
switches apply only to the connection that is being opened, and
do not affect the global FTP PASSIVE-MODE setting.
/ACTIVE
Opens the connection in active mode. Use this switch if the
server does not support passive mode, or use the command SET FTP
PASSIVE-MODE OFF.
/NOINIT
Added in C-Kermit 8.0.201. Tells C-Kermit not to send REST,
STRU, FEAT, and MODE commands to the server when the connection
is opened, since these have been reported to cause confusion in
certain servers.
When a username or password is missing, a prompt is issued at the
controlling terminal and you must type the response; the response can
not be scripted. Use the switches to avoid prompts, or one of the
secure authentication methods described in the next section, or see
[236]SET FTP AUTOLOGIN and the [237]FTP USER and similar commands
described later in this section.
Examples:
ftp open kermit.columbia.edu /anonymous ; Open and log in anonymously
ftp kermit.columbia.edu /anonymous ; The OPEN keyword can be omitted
ftp xyzcorp.com ; Open and maybe prompt for username
ftp xyzcorp.com /user:olga ; Open and log in as olga
ftp testing.abccorp.com 449 ; Specify a special TCP port number
ftp testing.abccorp.com /user:olaf /password:secret 449
The FTP OPEN command succeeds if a connection was opened to the server
(even if the given username and password were not valid) and fails
otherwise (see [238]Section 3.8 for details).
When your FTP session is complete, you can terminate it as follows:
FTP BYE
Closes the FTP connection if one was open. The FTP prefix can be
omitted if no other connection is open at the same time (see
[239]Section 3.8 for details). If a connection log is active, an
FTP record is written to it. If Kermit was started with the -9
command-line option or with its FTP command-line personality,
and the -S (Stay) option was not given, AND there is no other
active connection, the FTP BYE command also exits, just as it
does on a regular FTP client. Synonyms: FTP CLOSE, FTP QUIT (but
if the FTP prefix is omitted from QUIT, this becomes the regular
Kermit QUIT command, which is equivalent to EXIT; i.e. it closes
the connection and exits from Kermit).
The following commands can be used to achieve greater control over the
connection and login process:
SET FTP ANONYMOUS-PASSWORD text
Allows you to choose the password text to be sent automatically
by Kermit when you open an FTP connection with the /ANONYMOUS
switch.
SET FTP AUTOLOGIN { ON, OFF }
If you give this command prior to opening an FTP connection, it
controls whether Kermit tries to log you in automatically as
part of the connection process. Normally ON, which means the
username and password are sent automatically (and prompted for
if they are not yet known). When OFF, FTP OPEN connects to the
server without logging in. OFF is equivalent to the -n
command-line option when using Kermit's FTP command-line
personality.
FTP USER name [ password [ account ] ]
Used to log in to an FTP server to which a connection has been
made without autologin, or when autologin failed. If the
password is furnished on the command line, it is used; otherwise
you are prompted for a password. An account may also be
furnished if required by the server; it is not required by
Kermit and is not prompted for if omitted. Synonyms: USER, FTP
LOGIN.
FTP ACCOUNT text
Sends an account name to a server that supports accounts. If the
server does not support accounts, an error response occurs. If
the server does support accounts, the account is accepted if it
is valid and rejected if it is not. The account might be used
for charging purposes or it might be a secondary password, or it
might be used for any other purpose, such as an access password
for a particular disk. Servers that support accounts might or
might not allow or require the account to be sent prior to
login; usually it is sent after login, if at all. Synonym:
ACCOUNT.
Example:
set ftp autologin off ; One thing at a time please
ftp xyzcorp.com ; Try to make the connection
if fail exit 1 FTP connection failed ; Check that it was made
ftp user olga secret ; Now log in to the server
if fail exit 1 FTP login failed ; Check that it worked
ftp account 103896854 ; Login OK - send account
if fail echo WARNING - FTP ACCT failed ; Warn if problem
... ; (have session here)
bye ; Log out and disconnect
The following commands are used to control or get information about the
FTP connection. Any particular FTP server does not necessarily support
all of them.
FTP RESET
Terminates a user session but leaves the connection open,
allowing a new login via FTP USER.
FTP IDLE [ number ]
Most FTP servers automatically log you out and and disconnect
your session if there has been no activity for a certain amount
of time. Use this command to ask the server to set its idle
limit to the given number of seconds. Omit the number to ask the
server to inform you of its current idle limit.
FTP STATUS [ filename ]
Asks the FTP server to send information about the current
session. The result is a free-format report that might include
server identification, username and login time, FTP protocol
settings, and file-transfer statistics. If a filename is given,
the server is supposed to send detailed information about the
file.
FTP SYSTEM
Asks the FTP server to identify its operating system (Listed in
Internet Assigned Numbers, Operating System Names). Examples:
UNIX, VMS, VM/CMS, WINDOWS-NT. Unfortunately many variations are
allowed (e.g. LINUX-2.0, LINUX-2.2, FREEBSD, ULTRIX, etc,
instead of UNIX; WINDOWS-NT-3, WINDOWS-NT-3.5, WINDOWS-NT-3.51,
WINDOWS-NT-4, etc). The report might also include other
information like "Type L8", "Type I", or "Type A", indicating
the file-transfer mode.
FTP HELP [ keyword [ keyword [ ... ] ]
Asks the server to list the commands it supports. The response
is usually cryptic, listing FTP command mnemonics, not the
commands used by the client (since the server has no way of
knowing anything about the client's user interface). For
example, the PUT command is STOR in FTP protocol. If a keyword
is given, which should be an FTP protocol command,
slightly-more- detailed help is given about the corresponding
command (if the FTP server supports this feature). Examples:
"ftp help", "ftp help stor".
FTP SITE text
(Advanced) Sends an FTP SITE (site-specific) command. Usually
this means that the FTP server is asked to run an external
command with the given arguments. You might be able to find out
what SITE commands are available by sending "ftp help site" to
the server, but in general the availability of and response to
SITE commands is (not surprisingly) site specific.
FTP QUOTE text
(Advanced) Sends an FTP command in FTP protocol format. Use this
command to send commands to the server that the FTP client might
not know about.
SHOW FTP
Lists client (Kermit) FTP settings and information. Also SHOW
CONNECTION, SHOW COMMUNICATIONS.
HELP FTP [ keyword ]
Asks Kermit to list and describe its built-in FTP commands.
HELP SET FTP [ keyword ]
Asks Kermit to list and describe its built-in SET FTP commands.
[ [240]Top ] [ [241]FTP Top ] [ [242]C-Kermit Home ] [ [243]Kermit Home
]
3.1.5. Making Proxy Connections
The simplist form of firewall traversal is the HTTP CONNECT command.
The CONNECT command was implemented to allow a public web server which
usually resides on the boundary between the public and private networks
to forward HTTP requests from clients on the private network to public
web sites. In order to allow secure web connections to be established,
the CONNECT command works by authenticating the client with a
username/password and then establishing a tunnel to the desired host.
Many web servers support the CONNECT command and it can be configured
to allow outgoing connections to authenticated user to any TCP/IP
hostname/port combination accessible to the web server. The limitations
of HTTP CONNECT is that it can only be used for outgoing connections
for protocols that are implemented using TCP/IP. Protocols such as
Kerberos authentication that use UDP/IP cannot be tunneled using HTTP
CONNECT.
Kermit provides support for the use of HTTP CONNECT proxy services with
the command:
SET TCP HTTP-PROXY [/USER:username /PASSWORD:password] hostname/ip-address[:po
rt]
When a port is not specified the default port configured on the HTTP
server is used. This is frequently port 443. When a hostname is
specified, it is resolved using the DNS available to the web server.
3.2. Making Secure FTP Connections
Also see: [244]Accessing IBM Information Exchange with Kermit.
In the previous section, you can see several examples of traditional
insecure authentication: username and password sent across the network
in clear text. Of course this is bad practice on at least two counts:
(1) storing passwords in files (such as script files) gives access to
the target systems to anybody who can obtain read access to your
scripts; and (2) sending this information over the network leaves it
open to interception by network sniffers or compromised hosts.
Because of the increasing need for security on the Internet, FTP
servers are beginning to appear that offer secure forms of
authentication, in which no information is sent over the network that
would allow anyone who intercepts it to usurp your identity and gain
your access rights.
Kermit provides an equivalent form of FTP security for each type of
IETF standard security implemented in Telnet. These include
GSSAPI-KERBEROS5, KERBEROS4, Secure Remote Password (SRP), and
Transport Layer Security (SSL and TLS). It does not presently include
SSL tunneling nor any form of SSH v1 or v2. When Kermit is built with
the necessary libraries, secure FTP connections are attempted by
default, in which all connections are authenticated and the command and
data channels are private.
The use of authentication and encryption for FTP connections can be
adjusted with the commands listed below, which are available only if
your version of Kermit was built with the corresponding security
options and libraries:
SET FTP AUTHTYPE { AUTOMATIC, GSSAPI-KRB5, KERBEROS4, SRP, SSL, TLS }
Specifies an ordered list of authentication methods to be
attempted when AUTOAUTHENTICATION is ON. The default list is:
GSSAPI-KRB5, SRP, KERBEROS_V4, TLS, SSL. If none of the selected
methods are supported by the server, an insecure login is used
as a fallback. Note, by the way, that SSL or TLS can be used to
secure an anonymous connection.
SET FTP AUTOAUTHENTICATION { ON, OFF }
Tells whether authentication should be negotiated by the FTP
OPEN command. Default is ON. Use SET FTP AUTOAUTHENTICATION OFF
to force a clear-text, unencrypted connection to FTP servers
(such as the one at the Kermit FTP site) that normally would try
to negotiate secure authentication and encryption.
SET FTP AUTOENCRYPTION { ON, OFF }
Tells whether encryption (privacy) should be negotiated by the
FTP OPEN command, which can happen only if secure authentication
is also negotiated. Default is ON.
SET FTP AUTOLOGIN { ON, OFF }
Tells Kermit whether to try logging in automatically when you
make an FTP connection, as opposed to letting you do it "by
hand" with the FTP USER command.
SET FTP COMMAND-PROTECTION-LEVEL { CLEAR, CONFIDENTIAL, PRIVATE, SAFE }
Determines the level of protection applied to the command
channel:
CLEAR Data is sent in plaintext and not protected against tampering.
CONFIDENTIAL Data is encrypted but not protected against tampering.
PRIVATE Data is encrypted and is protected against tampering.
SAFE Data is sent in plaintext but protected against tampering.
The default is PRIVATE.
SET FTP CREDENTIAL-FORWARDING { ON, OFF }
Tells whether end-user credentials are to be forwarded to the
server if supported by the authentication method (GSSAPI-KRB5
only). This is often required to allow access to distributed
file systems (e.g. AFS.)
SET FTP DATA-PROTECTION-LEVEL { CLEAR, CONFIDENTIAL, PRIVATE, SAFE }
Tells what level of protection is applied to subsequent data
channels. The meanings of the protection-level keywords are the
same as for SET FTP COMMAND-PROTECTION-LEVEL. The default is
PRIVATE.
SET FTP SRP CIPHER name
Specifies the cipher to be used for encryption when SRP
authentication is in use. The list of possible choices is
computed based on the capabilities of the local SRP library and
includes NONE plus zero or more of the following:
BLOWFISH_ECB CAST5_ECB DES_ECB DES3_ECB
BLOWFISH_CBC CAST5_CBC DES_CBC DES3_CBC
BLOWFISH_CFB64 CAST5_CFB64 DES_CFB64 DES3_CFB64
BLOWFISH_OFB64 CAST5_OFB64 DES_OFB64 DES3_OFB64
The default is DES3_ECB.
SET FTP SRP HASH name
Specifies the hash to be used for data protection when SRP
authentication is in use. The choices are MD5 and SHA. The
default is SHA.
Command-line options:
-k name
Specifies the realm to be used with Kerberos 4 authentication (=
SET AUTH K4 REALM name).
-f
Enables forwarding of Kerberos 5 credentials to the host when
using GSSAPI authentication (= SET AUTH K5 FORWARDABLE ON).
-x
Enables autoencryption (= SET FTP AUTOENCRYPTION ON).
-c cipher
Specifies the kind of cipher to be used for encryption with SRP
authentication. Equivalent to SET FTP SRP CIPHER, with the same
choices. If this option is not given, CAST5_CBC is used.
-H hash
Specifies the hash to be used for encryption with SRP
authentication. Equivalent to SET FTP SRP HASH, with the same
choices. If this option is not given, SHA is used.
-z debug
Turns on SSL/TLS debugging.
-z secure
Requires secure connection.
-z certsok
Says to accept all certificates without checking validity.
-z verify=n
Sets certificate verification mode to the given number, n:
0 = no verification
1 = verify certificate if presented
2 = require verification of certificate
-z cert=filename
Specifies a file containing a client certificate to be presented
to the FTP server.
-z key=filename
Specifies a file containing a private key matching the client
certificate.
-z !krb4
(nokrb4) Disables the use of Kerberos 4.
-z !gss
-z nogss
Disables the use of GSSAPI - Kerberos 5.
-z !srp
-z nosrp
Disables use of SRP.
-z !ssl
-z nossl
Disables the use of SSL.
-z !tls
-z notls
Disables the use of TLS.
Caution: If your FTP connection is secured via AUTH TLS, it is not
possible to interrupt a file transfer. This is a limitation of all
known FTP servers that support AUTH TLS.
Note that when using certain security methods, such as SSL or TLS, you
may be prompted to confirm or verify certain actions or conditions, for
example, whether to accept self-signed certificates. This can interfere
with unattended operation of scripts; see [245]Section 3.10.
[ [246]Top ] [ [247]FTP Top ] [ [248]C-Kermit Home ] [ [249]Kermit Home
]
3.3. Setting FTP Preferences
FTP preferences can be set globally and persistently with the commands
in the following sections; many of these can also be overridden on a
per-command basis with switches that have the same name.
3.3.1. Logs, Messages, and Other Feedback
You can control the amount of feedback received from your FTP session
with the commands in this section. First, you can create a log of your
FTP transfers with the following commands:
SET TRANSACTION-LOG { VERBOSE, FTP, BRIEF }
Selects the log format. VERBOSE is the default, and is described
in [250]the manual. FTP chooses a WU-FTPD format, the same as is
used by the popular FTP server. BRIEF creates per-file records
in comma-separated-list format. For greater detail, see
[251]Section 4.17 of the [252]C-Kermit 7.0 Update Notes.
LOG TRANSACTIONS filename
Records FTP (or Kermit, or any other protocol) uploads and
downloads in the given file using the format selected by the
most recent SET TRANSACTION-LOG command, if any, or else the
default format.
FTP screen messages and displays are controlled by the following
commands:
SET TRANSFER DISPLAY { FULLSCREEN, CRT, SERIAL, BRIEF, NONE, OFF }
FTP transfers use Kermit's normal file-transfer display styles.
Use this command to choose the desired format; the default on
most platforms is FULLSCREEN. The display is automatically
disabled if Kermit is running in the background or in batch.
BRIEF is always used for command-line initiated transfers
(unless suppressed by -q). While a file-transfer is in progress,
you can interrupt it in the normal Kermit way by typing one of
the following keys or key combinations:
X - Cancel current file but go on to the next one (if any).
Z - Cancel the entire transfer. Ctrl-L or Ctrl-W - Refresh
the file-transfer display (if any).
SET FTP DISPLAY { FULLSCREEN, CRT, SERIAL, BRIEF, NONE, OFF }
Like SET TRANSFER DISPLAY, but applies only to FTP connections,
and does not affect Kermit- or other protocol file transfers.
SET QUIET { ON, OFF }
This command applies to Kermit in general, not just FTP. OFF by
default; when ON, it suppresses most messages from most commands
as well as the file-transfer display.
SET FTP PROGRESS-MESSAGES { ON, OFF }
Tells whether Kermit should print locally-generated feedback
messages for each non-file-transfer command. ON by default.
SET FTP VERBOSE-MODE { ON, OFF }
Tells whether to display all responses from the FTP server. OFF
by default. This shows all responses to all commands, except
when the file-transfer display is active, and unless you have
SET QUIET ON. When OFF, responses are shown only for commands
such as FTP PWD whose purpose is to display a response.
SET FTP DEBUG { ON, OFF }
Tells whether local client debugging information should be
displayed. OFF by default. When ON, the commands that are sent
to the server are shown, as well as its responses (even if
VERBOSE-MODE is OFF), plus additional informational messages are
printed regarding the progress of secure operations. Also, the
temporary file created by the [253]MGET command is not deleted
so you can see what's in it.
Set all of these to OFF when silent running is desired.
3.3.2. Operational Preferences
FTP DISABLE new-protocol-feature-name
FTP ENABLE new-protocol-feature-name
Explained in [254]Section 3.11.
SET FTP AUTOLOGIN { ON, OFF }
If you give this command prior to opening an FTP connection, it
controls whether Kermit tries to log you in automatically as
part of the connection process. Normally ON, which means the
username and password are sent automatically (and prompted for
if they are not yet known). When OFF, FTP OPEN connects to the
server without logging in. OFF is equivalent to the -n
command-line option when using Kermit's FTP command-line
personality. See [255]Section 3.1.4 for usage.
SET FTP PASSIVE-MODE { ON, OFF }
ON by default, to avoid random TCP port assignment for data
connections, which can prevent FTP protocol from working through
firewalls and network address translators (for more on these
topics, see the [256]Kermit security reference. Set to OFF in
case the FTP server does not support passive mode, or in case
the client has problems with it (it has been observed, for
example, that when using passive mode, the SCO XENIX 2.3.4
TCP/IP stack hangs in the connect() call forever). Synonyms:
PASSIVE [ ON ], PASSIVE OFF, PASV [ ON ], PASV OFF.
SET FTP SEND-PORT-COMMANDS { ON, OFF }
This command determines whether the FTP client sends a new PORT
command to the server when accepting incoming data connections
(as when not using passive mode.) When PASSIVE-MODE is OFF and
SET SEND-PORT is OFF, the port that was originally specified is
reused. This is the default behavior for normal FTP clients but
it is not compatible with many firewalls.
SET FTP CHARACTER-SET-TRANSLATION { ON, OFF }
Whether to translate character sets when transferring files with
FTP (explained in [257]Section 3.7). OFF by default.
SET FTP SERVER-CHARACTER-SET name
Tells Kermit the character set used by the FTP server, UTF-8 by
default ([258]Section 3.7).
SET FTP SERVER-TIME-OFFSET delta-time
Tells Kermit to apply the given [259]delta time to file
timestamps provided by the server for its files; for use when
(for example) the server does not have its timezone set
correctly.
SET FTP ERROR-ACTION { PROCEED, QUIT }
When transferring a group of files with FTP, and an error occurs
with one of the files, Kermit normally goes on the next file.
Use SET FTP ERROR-ACTION to QUIT to make Kermit stop the
transfer immediately and fail if an error occurs with any single
file in the group. Example: you have given Kermit a list of
files to send, and one of the files can not be found, or read
permission is denied. Note that canceling a file by typing 'X'
during transfer is not considered an error (if you want to
cancel the entire transfer, type 'Z' or Ctrl-C).
SET FTP PERMISSIONS { AUTO, ON, OFF }
When uploading files with PUT or MPUT, this tells whether Kermit
should send each file's permissions. The default is OFF, which
means not to send permissions, in which case the uploaded file's
permissions are set by the FTP server according to its own
criteria. ON means to send them, AUTO means to send them only if
the client (Kermit) and server are on like platforms (e.g. both
UNIX). This command has no effect when downloading, since the
FTP protocol does not include a way for the server to inform the
client of a file's permissions. Also see [260]FTP PUT
/PERMISSIONS. Note that setting permissions after uploading is
likely to work (correctly or at all) only when the client and
server platforms are alike (e.g. both of them are some form of
UNIX). Also note that Windows files don't have permissions. Also
see [261]FTP CHMOD.
SET FTP DATES { ON, OFF }
When downloading files with GET or MGET, this tells whether
Kermit should try to set the received file's date from the
server's date. FTP DATES is ON by default. Note, however, that
FTP protocol does not allow date preservation when uploading. So
at best, SET FTP DATES ON can work only when downloading, and
then only when the server agrees to furnish file dates.
SET FTP FILENAMES { AUTO, CONVERTED, LITERAL }
When uploading (sending) files, this tells whether to convert
outbound filenames to "common form". This means allowing only
one period in a name, uppercasing any lowercase letters,
replacing spaces by underscores, etc. AUTOMATIC is the default,
meaning LITERAL when client and server are the same type of
system (e.g. UNIX) and CONVERTED otherwise. Special case: if the
setting is AUTOMATIC and the client is not UNIX and the server
identifies itself as UNIX, Kermit uses a less-strict form of
conversion, in which lowercase letters are not uppercased and
the filename can contain any number of periods, but spaces are
still converted to underscore. When receiving, conversion
generally means to change all-uppercase names to lowercase and
spaces to underscore.
SET FTP UNIQUE-SERVER-NAMES { ON, OFF }
Applies only to uploads. Tells the server to create new, unique
names for incoming files that have the same names as existing
files. OFF by default, in which case the server overwrites
existing files with new files of the same name. When ON, the
server uses its own built-in method for creating new names for
incoming files; for example, appending a period (.) and a number
to the name. CAUTION: Use this option only if you do not need to
refer to the file after it is uploaded, since FTP protocol
provides no mechanism for the client to find out what name was
assigned by the server.
SET FTP COLLISION { ... }
When downloading, what to do if an incoming file has the same
name as an existing file. Options are the same as for SET FILE
COLLISION. If this command is not given, Kermit's regular FILE
COLLISION setting is used. If this command is given, it
overrides the FILE COLLISION setting for FTP transfers only. See
[262]Section 3.6.2 for details.
SET FTP TYPE { TEXT, BINARY, TENEX }
Changes the default transfer mode. When sending (uploading)
files, this command has no effect unless you disable automatic
text/binary mode switching ([263]Section 4) with SET FILE SCAN
OFF or SET TRANSFER MODE MANUAL. When receiving (downloading)
files, this command establishes the transfer mode to be used
when a filename does not match any of Kermit's text or binary
filename patterns, unless you use SET FTP GET-FILETYPE-SWITCHING
or SET TRANSFER MODE MANUAL to disable automatic switching, in
which case, this command establishes the transfer mode for all
downloaded files. In all cases, however, the FTP TYPE can be
overridden in any GET or PUT command by including a /TEXT
(/ASCII), /BINARY, or /TENEX switch. The FTP TYPE is independent
of the Kermit FILE TYPE setting. TENEX is used for sending 8-bit
binary files to 36-bit platforms such as TOPS-10, TOPS-20, and
TENEX, and getting them back again. Synonym: ASCII = TEXT. Note:
there is also an FTP TYPE command, which does what SET FTP TYPE
does but also sends a TYPE command to the server immediately if
the given type is different from the current one.
If you want want specific FTP preference settings to be in effect for
all your Kermit FTP sessions, put the desired SET FTP commands in your
Kermit customization file (~/.mykermrc in UNIX, K95CUSTOM.INI in
Windows).
[ [264]Top ] [ [265]FTP Top ] [ [266]C-Kermit Home ] [ [267]Kermit Home
]
3.4. Managing Directories and Files
In Kermit, commands for directory and file management can refer to:
* The local computer
* A remote computer when you have a connection to a Kermit server or
IKSD.
* A remote computer when you have a connection to an FTP server.
(There can also be an HTTP connection, but the commands in this section
don't apply to HTTP connections.)
Thus in general, each such command comes in three forms:
1. With no prefix in C-Kermit 8.0.200, it refers to the local computer
(CD, DIR, etc). In C-Kermit 8.0.201 and later, however, the "locus"
switches to automatically to the remote FTP server when you make an
FTP connection (see the SET LOCUS description [268]Section 7); thus
C-Kermit 8.0.201 acts almost exactly like a regular FTP client when
it has an FTP connection, yet still acts like itself on other kinds
of connections.
2. With the REMOTE prefix, it is for a Kermit server (REMOTE CD,
REMOTE DIR).
3. With the FTP prefix, it's for an FTP server (FTP CD, FTP DIR).
4. Also see [269]Section 3.8, which explains "R-commands" and
"L-commands".
Kermit's FTP file and directory management commands are as follows.
When an R-command is included in the Synonyms list, be sure to read
[270]Section 3.8 about rules for use of R-commands.
FTP CD [ directory ]
Tells the FTP server to change its default (working) directory
to the one given, which usually must be expressed in the syntax
of the server platform (UNIX, VMS, etc). If the directory is not
specified, the result depends on the FTP server -- it might
complain that the command is illegal, or it might change to your
original login directory. Synonyms: FTP CWD (Change Wording
Directory); RCD.
FTP CDUP
Tells the FTP server to change its default (working) directory
to the parent directory of its current one (equivalent to
"cd .." in UNIX, or "cd [-]" in VMS). Synonyms: RCDUP, FTP UP.
FTP PWD
Asks the FTP server to report ("print") its current working
directory. Synonym: RPWD.
FTP MKDIR directory
Asks the FTP server to create the directory whose name is given.
In general, the name must be in the syntax of the server's file
system, and it must be either absolute (a full pathname) or
relative to the server's current (working) directory. This
command fails if the directory can't be created for any reason,
including that it exists already. Synonym: RMKDIR.
FTP RMDIR directory
Asks the FTP server to remove the directory whose name is given.
The rules are the same as for MKDIR, plus in most cases, the
server will not remove any directory unless it is empty.
Synonym: RRMDIR.
FTP DIRECTORY [ filespec ] [ redirectors ]
Tells the FTP server to send a directory listing of the
specified files. If no filespec is given, the server lists all
files in its current working directory. The results are in
whatever format the server chooses to send them. You can use
UNIX-like redirectors to send the listing to a file or a
pipeline, exactly as with the regular Kermit client/server
REMOTE DIRECTORY command ([271]Using C-Kermit, Chapter 11).
Synonym: RDIRECTORY. Examples:
ftp dir ; Show listing of all files on screen
ftp dir *.txt ; List *.txt files on screen
ftp dir *.txt > somefile ; Put listing in somefile
ftp dir *.txt >> somefile ; Append listing to somefile
ftp dir *.txt | sort > somefile ; Put sorted listing in somefile
ftp dir | more ; Runs list through "more"
ftp dir | sort | more ; Runs list through "sort" and "more"
FTP VDIRECTORY [ filespec ] [ redirectors ]
"Verbose" directory. This is an alternative FTP DIRECTORY
command primarily for use with DECSYSTEM-20 (TOPS-20) FTP
servers, which send only filenames when given a DIRECTORY
command; the VDIRECTORY command makes them also send file sizes,
dates, and attributes.
FTP CHECK filespec
Asks the FTP server whether the given file exists or, if the
filespec contains wildcards, if any files match, and this
command succeeds or fails accordingly.
FTP MODTIME filename
Asks the FTP server, via the not-yet-standard FTP MDTM command,
to send the modification date and time of the given file. The
response should be a numeric string in the format:
yyyymmddhhmmssxxxxx... where yyyy is the year, mm is the month,
dd is the day, hh is the hour (0-23), mm is the minute, ss is
the second, and xxx... is the optional fraction of the second (0
or more digits). The date and time is expressed in UTC (GMT,
Zulu, Zero-Meridian). The result is available programmatically
in the [272]\v(ftp_message) variable, and is understandable by
Kermit's date-time switches and functions. For example, suppose
we want to upload all local files that are newer than a
particular file on the server:
C-Kermit> ftp modtime signpost
C-Kermit> echo \v(ftp_message)
20010807113542.014
C-Kermit> ftp mput /after:\v(ftp_message)GMT *
Note that we must append "GMT" to the date-time string to let
the /AFTER switch know the time is GMT rather than local.
FTP SIZE filename
Asks the FTP server to send the size (in bytes) of the given
file. The result might vary depending on whether the current FTP
TYPE is binary or text ("ascii"). For a reliable byte count, do
FTP TYPE BINARY first. The result is available programmatically
in the [273]\v(ftp_message) variable.
FTP CHMOD permissions filename
Tells the FTP server to set the permissions (protection) of the
given file to the ones given. The permissions and filename must
be given in whatever syntax is required by the server. Example
(for a UNIX-based FTP server):
ftp chmod 664 oofa.txt
Not all servers support this command. For non-UNIX-based
servers, you might need to use FTP QUOTE or FTP SITE and the
appropriate platform-specific FTP server command.
FTP UMASK [ number ]
This command is probably specific to UNIX-based servers; it sets
the UNIX "umask", which is the default permissions mask for new
(in this case, incoming) files. Crudely put, the UNIX umask is
an octal representation of a binary number in in which a 1 bit
stands for a permission bit that must be 0, and a 0 bit stands
for a permission bit that can be 0 or 1 depending on other
factors, such as the permissions of the parent directory.
Example: "umask 007" requires that new files are created without
read/write/execute world permission. If the number is not
specified, the server's current umask is reported.
FTP RENAME filename newname
Asks the FTP server to rename the file whose name is "filename"
to "newname". Works only for one file; can not be used with
wildcards. The server's interpretation of "newname" can vary (in
some cases it must be a filename, in others perhaps it can also
be a directory name, in which case if the filename denote a
regular file, the file might be moved to the given directory).
Some servers might allow files to be renamed ("moved") between
physical disks or partitions, others might not. Synonym:
RRENAME.
FTP DELETE [ switches ] filespec [ filespec [ ... ] ]
Tells the FTP server to delete the file or files listed. Each
file specification may, but need not, contain wildcard
characters to match multiple files. File specifications and
wildcard syntax must be those of the server. Any file
specifications that contain spaces must be enclosed in braces or
doublequotes. FTP DELETE switches are:
/ERROR-ACTION: /FILENAMES: /NOBACKUPFILES /QUIET
/EXCEPT: /LARGER-THAN: /NODOTFILES /NOPAGE
/PAGE /RECURSIVE /SMALLER-THAN:
When used with FTP DELETE, the /RECURSIVE switch deletes files
but not directories, and furthermore depends on the server
providing recursive file lists, which is not the normal
behavior. For further details, see the descriptions of these
switches in [274]Section 3.6. Synonyms: FTP MDELETE (Kermit
makes no distinction between DELETE and MDELETE); RDELETE.
FTP TYPE { TEXT, BINARY, TENEX }
Tells the FTP server to change its file-transfer type to the one
given, immediately. See [275]SET FTP TYPE for details.
[ [276]Top ] [ [277]FTP Top ] [ [278]C-Kermit Home ] [ [279]Kermit Home
]
3.5. Uploading Files With FTP
Uploading means sending files from the client (Kermit) to the FTP
server. The basic command for uploading files with FTP is PUT:
FTP PUT [ switches ] [ filespec [ as-name ] ]
Uploads (sends) the file or files that match the file
specification, which may include wildcards, to the server. If no
filespec is given, the names of files to send are taken from the
/LISTFILE: file, if any, otherwise from the SEND-LIST, if any.
Unless you go out of your way to prevent it, Kermit determines
the transfer mode (text or binary) for each file automatically,
and switches automatically on a per-file basis. If an as-name is
given, the file is sent under that name instead of its own (if
an as-name is given with a wildcard filespec, the result is a
bit more complicated, and is explained later in this section).
Unlike normal FTP clients, Kermit does not prompt you by default (or at
all) for each file; it just sends them, just as it does with Kermit
protocol. The filespec can be a literal filename or a Kermit pattern,
described in:
[280]http://www.columbia.edu/kermit/ckermit70.html#x4.9
Kermit patterns are equivalent to C-Shell patterns and provide a fair
amount of flexibility in selecting which files to send, which is
augmented by the file-selection switches presented in [281]Section
3.5.1.
FTP MPUT [ switches ] filespec [ filespec [ ... ] ]
FTP MPUT is just like FTP PUT except it allows you to give more
than one file specification, and it does not allow an as-name in
the file list. However, as-names can be given to either PUT or
MPUT with the /AS-NAME: switch.
If a PUT or MPUT command results in one file being uploaded, it
succeeds if the file is uploaded completely and fails otherwise. If
more than one file is selected for upload, success or failure depends
on the [282]FTP ERROR-ACTION setting; if it is PROCEED (the default
setting), then the [M]PUT command succeeds if at least one of the files
was completely uploaded, and fails otherwise, If FTP ERROR-ACTION is
QUIT, the [M]PUT command succeeds if all selected files were uploaded
successfully, and fails if any file failed.
FTP uploads may be interrupted just like Kermit uploads. While the
transfer is in progress, type:
X to interrupt the current file and go on to the next file.
Z to cancel the current file and all remaining files.
^C (Control-C): Like Z, but might act more quickly.
MPUT may be used as in regular FTP clients, but it is not required to
send multiple files; in Kermit it is required only if you want to give
multiple file specifications. Examples:
ftp put oofa.txt ; Send a single file oofa.txt
ftp put oofa.txt budget.txt ; Send single file oofa.txt as budget.txt
ftp put *.txt ; Send all *.txt files
ftp mput *.txt ; Send all *.txt files (same as "put *.txt")
ftp mput *.txt foo.bar ; Send all *.txt files plus foo.bar
The distinction between PUT and MPUT is important only when more than
one filespec is given, just like the distinction between Kermit SEND
and MSEND:
ftp put oofa.txt budget.txt ; Send oofa.txt AS budget.txt
ftp mput oofa.txt budget.txt ; Send oofa.txt AND budget.txt
If the source file specification includes any path segments, for
example:
put /tmp/oofa.txt
put subdir/another/andanother/oofa.txt
the path portion is stripped from the filename that is sent to the
server. However, if an as-name contains a path, it is retained.
Examples:
ftp put /usr/doc/oofa.txt ; Send as "oofa.txt".
ftp put oofa.txt /tmp/oofa.txt ; Send as "/tmp/oofa.txt"
The latter example sends the file oofa.txt from your current local
directory to the server's /tmp directory. This works only if the server
uses the same directory notation that you used in the as-name AND the
given directory already exists on the server AND if you have write
access to it.
Use caution when uploading from a case-sensitive file system, such as
UNIX, to a file system that is not case sensitive, such as Windows or
VMS. If you have two files in UNIX, AA and aa and upload both of them,
the second one will overwrite the first. The only way around this
provided by FTP protocol is its "unique server names" feature (SET FTP
UNIQUE-SERVER-NAMES or the /UNIQUE switch described below).
3.5.1. FTP PUT Switches
FTP PUT and MPUT are similar in format and behavior to the regular
Kermit SEND and MSEND commands, and they allow most of the same
optional switches:
C-Kermit>ftp put ? Filename, or switch, one of the following:
/after: /larger-than: /rename-to:
/array: /listfile: /server-character-set:
/as-name: /local-character-set: /server-rename-to:
/before: /move-to: /simulate
/binary /nobackupfiles /smaller-than:
/command /nodotfiles /tenex
/delete /nofollowlinks /text
/dotfiles /not-after: /transparent
/error-action: /not-before: /type:
/except: /permissions: /update
/filenames: /quiet /unique-server-names
/filter: /recover
/followlinks /recursive
Since most of these switches are common to Kermit's SEND and MSEND
commands, they described only briefly here. For greater detail see:
[283]http://www.columbia.edu/kermit/ckermit70.html#x1.5 (explanation
of switches)
[284]http://www.columbia.edu/kermit/ckermit70.html#x4.7
(file-transfer switches)
First the file-selection switches:
/AFTER:date-time
/BEFORE:date-time
/NOT-AFTER:date-time
/NOT-BEFORE:date-time
Only send those files modified on or after or before the given
date and time. These switches can be combined to select files
modified between two date/times. Various date-time formats are
accepted; if the date-time contains spaces, it must be enclosed
in braces or doublequotes. See
[285]http://www.columbia.edu/kermit/ckermit70.html#x1.6 and
[286]Section 8.13 of this document for details about date-time
formats. Examples:
ftp put /after:{1 jan 2000 0:00:00} *
ftp put /after:-5days *
/LARGER-THAN:number
/SMALLER-THAN:number
Only send files larger (smaller) than the given number of bytes
(octets). These switches can be combined to select files in a
certain size range.
/TYPE:{TEXT,BINARY}
Only send files that are the given type, which is determined for
each file just before sending it by file scanning. BINARY
includes TENEX; if you have included a /TENEX switch, or
previously given a [SET] FTP TYPE TENEX command, binary files
are sent in TENEX, rather than BINARY mode.
/[NO]DOTFILES
[Don't] include files whose names begin with dot (.). By
default, such files are not included unless your filespec
explicitly mentions them.
/NOBACKUPFILES
Don't include files whose names end with .~nnn~, where nnn is a
number, e.g. oofa.txt.~27~. These are backup files created by
Kermit, EMACS, and other applications. By default, backup files
are included.
/NOFOLLOWLINKS
(UNIX only) Skip over symbolic links rather than following them
(default). This applies to wildcard and/or recursive [M]PUTs; if
a single filename is given, and it happens to be a symbolic
link, the file it points to is sent.
/FOLLOWLINKS
(UNIX only) Always follow (resolve) symbolic links, even in
wildcard or recursive [M]PUTs. Use with caution. Watch out for
circular links, endless loops, etc.
/EXCEPT:pattern
Exception list -- don't send files whose names match the given
pattern. See [287]Section 1.5.4 of the [288]C-Kermit 7.0 Update
Notes for details. If you want to exclude a directory from a
recursive [M]PUT, use /EXCEPT:{dirname/*}.
/RECURSIVE
Sends the desired files from the current (or given) directory,
plus all directories beneath it, including empty directories,
replicating the directory structure on the server. No special
capabilities are required in the server, but of course your
login ID on the server must have the appropriate access and
permission to create directories. Recursive PUTs work not only
between like platforms (e.g. UNIX to UNIX) but also between
unlike ones (e.g. UNIX to VMS or Windows), in which case
text-file format differences are handled by Kermit's automatic
text/binary mode switching ([289]Section 4) and character-set
translation ([290]Section 3.7). Synonym: /SUBDIRECTORIES.
/UPDATE
Send only files that have changed since last time ([291]Section
3.5.2).
/ARRAY:arrayname
The "file" to be sent is an array, or a segment of one, rather
than a real file. In this case the other selection switches
don't apply. The array contents are sent in text mode, and each
array element is treated as a line. Example:
ftp put /as-name:array.txt /array:&a
(or, to send a segment of the array, /array:&a[100:199]). If you
don't include an /AS-NAME, a name of "_array_x_" is used (where
x is the array letter). If you include this switch, most other
switches are meaningless and ignored.
/COMMAND
The "file" to be sent is the standard output of a command,
rather than a real file. It is sent in text or binary mode
according to the prevailing FTP TYPE, which can be overridden
with a /TEXT or /BINARY switch. Example: Example:
ftp put /command /as-name:{userlist} {finger | sort -r}
/LISTFILE:filename
Tells Kermit to obtain the list of files to be sent from the
file whose name is given. This file must contain one file
specification (which may be wild) per line. If the list includes
files from different directories, such as a recursive listing of
a directory tree, the paths are recreated on the server (if
possible) if you include the /RECURSIVE switch; otherwise all
the files are sent to the current directory on the server.
Now the other switches:
/AS-NAME:text
If a single file is being sent, send it with the given text as
its name. If multiple files are being sent, the text must be a
template that includes variables such as \v(filename),
\v(filenumber), \v(ntime), to allow dynamic creation of each
name. The same applies to the as-name field of the FTP PUT
command. If this switch is not included (and an as-name is not
included as the second filename to PUT), each file is sent with
its own name.
/BINARY
/TEXT
/TENEX
Forces this upload to take place in the given mode, regardless
of the current FTP TYPE setting, and without automatic
text/binary switching. /ASCII is a synonym for /TEXT.
/FILTER:command
Specifies that the file(s) is/are to be passed through the given
command or pipeline on their way to the server. Example:
ftp put /binary /filter:{gzip -c \v(filename)} /as-name:\v(filename).gz *
/TRANSPARENT
/LOCAL-CHARACTER-SET:name
/SERVER-CHARACTER-SET:name
Character-set translation for text files, explained in
[292]Section 3.7.
/ERROR-ACTION:{PROCEED,QUIT}
Overrides the prevailing [293]FTP ERROR-ACTION for the duration
of this PUT or MPUT command only.
/RECOVER
Resume an interrupted transfer where from the point of
interruption (explained in [294]Section 3.5.2). Synonym:
/RESTART.
/DELETE
Tells Kermit to delete each source file immediately after, and
only if, it has been uploaded completely and successfully. This,
in effect, moves the file from the client to the server.
/MOVE-TO:directory
Tells Kermit to move each source file to the named local
directory after, and only if, it has been uploaded completely
and successfully.
/RENAME-TO:template
Tells Kermit to rename each (local) source file according to the
given template after, and only if, it has been uploaded
completely and successfully. The template works as in /AS-NAME.
/SERVER-RENAME-TO:template
Tells Kermit to ask the server to rename each file according to
the given template as soon as, and only if, it has been received
completely and successfully. The template works as in /AS-NAME.
Requires write and rename access on the server, so doesn't
usually work with (e.g.) anonymous uploads to public incoming
areas where the permissions don't allow renaming. Examples:
ftp mput /server-rename:\v(filename).ok *
Appends ".ok" to each filename on the server when it's
finished uploading.
ftp mput /as-name:\v(filename).tmp /server-rename:\v(filename) *
This is the reverse of the previous example; it uses a
temporary name while uploading is in progress and reverts
the file to its real name when uploading is complete.
ftp mput /as-name:\v(filename)
/server-rename:../final/\v(filename) *
Moves the file from the working directory to a final
directory when the upload is complete, but in this case
you have to know the pathname syntax of the server. If the
rename fails, the [M]PUT command fails according to the
[295]FTP ERROR-ACTION selection.
/FILENAMES:{AUTOMATIC,CONVERTED,LITERAL}
Overrides the [296]FTP FILENAMES setting for this upload only.
/PERMISSIONS:{ON,OFF}
Overrides the [297]FTP PERMISSIONS setting for this upload only.
/UNIQUE
Tells Kermit to tell the server to give [298]unique names to
incoming files that would otherwise overwrite existing files
that have the same name. This switch conflicts with /UPDATE,
/RECOVER, /PERMISSIONS, and /SERVER-RENAME since the client has
no way of knowing the name assigned by the server.
/QUIET
Don't display file-transfer progress or statistics.
/SIMULATE
Shows which files would be sent without actually sending them.
Useful (for example) with /UPDATE (next section). The results
are shown in the file-transfer display (if it is not disabled)
and in the transaction log (if one is active). Hint: use SET
TRANSFER DISPLAY BRIEF.
3.5.2. Update Mode
When you include the /UPDATE switch, this means to skip sending any
file that already exists on the server if the local file's modification
date/time is not later than that of the corresponding file on the
server. Here is a typical application for update mode: Suppose that on
Computer A, you maintain a large set of files (say, a collection of Web
pages and graphics images, or the source files for a software
application), and you need to keep a parallel copy on another Computer,
B. Of course you could upload the entire collection every day:
cd source-directory
ftp computerb.xyzcorp.com
( authentication details... )
ftp cd target-directory
ftp put [ switches ] *
But if the total size is large or the network slow, this would be
unnecessarily time-consuming. Worse, if other users or sites had to
update whenever new files appeared in B's directory, this would cause
them unnecessary work. By including the /UPDATE switch:
ftp put /update [ other-switches ] *
only those files that changed since last time are uploaded. Here's how
it works. For each local file that is selected for uploading:
* The remote filename is determined in the normal way, according to
the [299]FTP FILENAMES setting, /FILENAMES switch, or the as-name,
if any.
* Kermit sends an MDTM (modification time) command for the
corresponding remote filename to the server.
* If the server does not understand the MDTM command, the file is
sent.
* If the server can't find a file with the given name, the file is
sent.
* If the local file's modification time is later than that of the
remote file, the file is sent.
* Otherwise -- the remote file exists but its modification time is
equal to or earlier than that of the local file -- the file is
skipped.
All time comparisons take place in Coordinated Universal Time
(UTC)([300]1), also known as GMT or Zulu time: Timezone 0; standard
time, without daylight savings.
WARNING: Some FTP servers, such as Novell NWFTPD.NLM, ignore or
misimplement the FTP specification and send local time rather than
UTC.
Update mode is useful only when always used in the same direction. When
you upload (PUT) a file with FTP, the destination file receives the
current timestamp on the server's computer, not the original file's
timestamp ([301]2). If you try to FTP PUT /UPDATE the same file again,
it will be skipped (as expected) since the remote copy is newer.
However, if you try to FTP GET /UPDATE the same file ([302]Section
3.6), it will be transferred for the same reason.
To check the availability of PUT /UPDATE on a particular connection,
issue an FTP MODTIME command for a file that is known to exist on the
server. If it succeeds, PUT /UPDATE should work and in that case, you
can run a procedure like the one above every day: the first time, it
sends all the files; after that, it sends only the ones that changed.
If a transaction log is active, a notation is included for any files
that are skipped.
Notes:
1. Why is Coordinated Universal Time abbreviated UTC? From the
[303]National Institute of Standards and Technology FAQ: "In 1970
the Coordinated Universal Time system was devised by an
international advisory group of technical experts within the
International Telecommunication Union (ITU). The ITU felt it was
best to designate a single abbreviation for use in all languages in
order to minimize confusion. Since unanimous agreement could not be
achieved on using either the English word order, CUT, or the French
word order, TUC, the acronym UTC was chosen as a compromise."
2. The Kermit FTP client is unusual in that, when downloading only, it
can set the received file's date from the file's date on the
server, but this should not affect the update feature. When
uploading to an FTP server, however, there is no mechanism for the
client to set the date of the uploaded file on the server.
3.5.3 Recovery
Suppose that while you are uploading a large file over a slow
connection, the connection is lost before the entire file is
transferred. With most FTP clients, you would have to start over, thus
resending the portion of the file that was sent already, and that is
already on the server. But Kermit's /RECOVER switch (Synonym: /RESTART)
lets you continue an interrupted transfer from the point of failure,
thus transferring only the part that wasn't sent already. The
prerequisites for recovery are:
* The transfer must be in BINARY mode, or else the client and server
must reside on like systems (e.g. both on some form of UNIX).
* The FTP server must support the SIZE command.
Here's how it works. When you include the /RECOVER switch:
* Kermit checks for conflicting switches, such as /UPDATE and
/UNIQUE; if /RECOVER is given with these switches an error occurs.
If /RECOVER is given in other circumstances where it could serve no
useful purpose (e.g. with arrays, pipes, or filters), it is
ignored.
If the switch is accepted, then for each selected file:
* If it is not binary (determined by scanning) and the client and
server are not on like platforms, recovery is canceled (the entire
file is sent). Otherwise:
* A SIZE command is sent for the file (using its remote name). If the
reply indicates the file was not found, or the SIZE command was not
understood, or any other kind of error, recovery is canceled.
Otherwise:
* A MDTM (modification time) command is sent for the file. If a valid
reply is received, and the modification time of the local file is
later than that of the remote file, recovery is canceled.
Otherwise:
* If the sizes of the two files are identical, the file is not sent.
Otherwise:
* Kermit seeks to the recovery spot in the local file, tells the
server to APPEND the data which is about to arrive to the remote
file, and then sends the data starting at the recovery point.
To safeguard file integrity, recovery is not attempted unless all the
preconditions are met. For the widest possible usefulness, APPEND is
used rather than RESTART. For stream transfers (the only kind that
Kermit supports) the results are the same.
By design, the /RECOVER switch can be included with any FTP PUT or MPUT
command, even if it specifies a group of files. This allows you to
resume an interrupted batch transfer from where it left off. The files
that were already completely sent are skipped, the file that was
interrupted is recovered, and the remaining files are uploaded.
By the way, it doesn't matter how the original partial file was
uploaded -- FTP, Kermit, Zmodem, etc: as long as the preconditions are
met, it can be recovered with FTP PUT /RECOVER, or for that matter also
using Kermit protocol and SEND /RECOVER.
A word of caution, however, when the original upload was in text mode
with character-set translation ([304]Section 3.7):
* If the original upload involved a translation from one single-byte
character set to another (e.g. Code Page 850 to Latin-1), recovery
is safe if you specify the same translations for the recovery. If
you don't, the resulting file will contain a mixture of character
sets.
* If the original upload involved a translation that changed the size
of the file (e.g. from an alphabetic Code Page or Latin Alphabet to
Unicode, or vice versa), recovery is NOT safe, even if you specify
the same translations.
Kermit has no way of knowing anything about the previous upload. As a
safeguard, an error occurs if you include /RECOVER and also specify a
character-set of UCS2 or UTF8, since recovery can't possibly work in
that situation. Otherwise, it's up to you to avoid unsafe recovery
operations.
[ [305]Top ] [ [306]FTP Top ] [ [307]C-Kermit Home ] [ [308]Kermit Home
]
3.6. Downloading Files With FTP
Although uploading files with Kermit's FTP client is just as easy and
flexible as sending files with Kermit protocol, the same is not always
true for downloading because FTP servers lack some of the capabilities
of a Kermit server:
* If you want to get more than one file, you have to use MGET, not
GET, since the underlying FTP protocol is different in the two
cases. Kermit can't "autodetect" which one you mean, as it can with
PUT and MPUT, since it can't be expected to know the wildcard
syntax of the remote platform and/or FTP server (the same is true
for all other FTP clients). To complicate matters, FTP protocol now
includes two underlying mechanisms (NLST and MLSD) for
accomplishing MGET operations and, as explained in [309]Section
3.11, the two behave differently.
* Automatic text-binary mode switching is not done by the server. It
can be done by the client (Kermit), but in this case it is not
based on a file scan (since there is no way for Kermit prescan a
server file), but rather on the filename, using C-Kermit 7.0
[310]filename patterns.
* Some options that are available with FTP PUT can not be used with
FTP [M]GET or don't work the same way:
/PERMISSIONS (FTP protocol has no mechanism for this).
/[NOT-]BEFORE, /[NOT-]AFTER (because of the timezone problem).
/RECOVER works only in binary mode. /RECURSIVE has limited
utility.
The commands for downloading are:
SET FILE DOWNLOAD-DIRECTORY [ directory ]
As with Kermit transfers, this command, if given, tells C-Kermit
where to store incoming files in the absence of a specific
as-name. If not given, incoming files are stored as indicated by
the as-name, if any, otherwise in the current directory, just as
with Kermit transfers. The more verbose transfer display formats
give the full pathname of each received file, and, in case you
have trouble finding a downloaded file afterwards, its full path
is also listed in the transaction log (if you kept one), and you
can also ask Kermit where it went with the [311]WHERE command.
SET FTP GET-FILETYPE-SWITCHING { ON, OFF }
ON by default, causing Kermit to switch automatically into text
or binary mode for each file based on whether its name matches a
text pattern or binary pattern. Set this OFF, or use a /TEXT,
/BINARY, or /TENEX switch to defeat this feature. Use SHOW
PATTERNS to see the current pattern list.
[ FTP ] GET [ switches ] filename [ as-name ]
Asks the server to send the given file, and if it comes, stores
it locally under the given as-name, if any, otherwise under its
original name (modified according to the selected filename
conversion option), in your download directory, if you have
specified one, otherwise in the directory indicated in the
as-name, if any, otherwise in your current directory. If you
accidentally use a wildcard in the filename ("get *.txt") the
server will reply with a message like "File not found" (unless
there is a file whose name actually is "*.txt"). If FTP
GET-FILETYPE-SWITCHING is ON, and in the absence of any GET
switches to override it, the file is transferred in binary mode
if it matches any of Kermit's binary name patterns, and in text
mode if it matches any of Kermit's text name patterns, and in
the prevailing FTP TYPE if it matches none of these patterns.
[ FTP ] MGET [ switches ] filespec [ filespec [ filespec [ ... ] ] ]
Like GET, but for multiple files. One or more file
specifications can be given, and any or all (or none) of them
can contain wildcards or can be directory names. The file list
may not include an as-name, but you can still give one with the
/AS-NAME: switch.
In both the FTP GET and MGET commands, any filenames that contain
spaces must be enclosed in braces or doublequotes (see [312]Section 5
for details).
FTP downloads may be interrupted just like Kermit transfers. While the
transfer is in progress, type:
* X to interrupt the current file and go on to the next file.
* Z (or Control-C) to cancel the current file and all remaining
files.
Before proceeding, a brief word about temporary files. In FTP protocol,
the MGET command works by requesting a file list from the server, and
then (internally) issuing a GET command (FTP RETR protocol directive)
for each file. The file list returned by the server can be any size at
all, so in case it is huge, we don't store it in memory; instead we put
it in a temporary file. For troubleshooting purposes, you should be
aware of two points:
1. The location of the temporary file is chosen according the TMP or
TEMP environment variables. If neither of these variables is
defined, you might need to define it. In case there is not enough
space on the indicated disk or partition for the server's file
list, you might need to either clean up the temporary area, or
redefine the environment variable to indicate a different area that
has sufficient space.
2. If you want to look at the list yourself, use SET FTP DEBUG ON.
This tells Kermit to (a) give you the full pathname of the
temporary file at the end of each MGET command, and (b) not to
delete it, as it normally does.
3.6.1. FTP GET Switches
The following switches are available with FTP GET and MGET:
/TEXT
Specifies a text-mode transfer. Overrides the global FTP TYPE
setting and filename pattern-matching for the duration of the
current command only, All files are downloaded in text mode.
Synonym: /ASCII.
/BINARY
Specifies a binary-mode transfer. Overrides the global FTP TYPE
setting and filename pattern-matching for the duration of the
current command only. All files are downloaded in binary mode.
/TENEX
Like /BINARY but specifies a special binary transfer mode to be
used when getting 8-bit binary files from a 36-bit platform such
as TOPS-10, TOPS-20, or TENEX. All files are downloaded in the
special binary mode.
/RECOVER
This instructs Kermit to try to recover an incomplete download
from the point of failure. Works only in binary mode, and only
if the server supports the (not-yet-standard) FTP "REST"
directive. See [313]Section 3.6.3 for details. Synonym:
/RESTART.
/FILENAMES:{CONVERTED,LITERAL}
Overrides the [314]FTP FILENAMES (filename conversion) setting
for this download only, forcing incoming filenames to be either
converted or taken literally.
/AS-NAME:text
For GET, this is equivalent to giving an as-name after the
filename. For MGET, this is the only way to specify alternative
names for the incoming files. With MGET, the /AS-NAME text
should (must) contain a Kermit variable, usually \v(filename) or
\v(filenumber). Example:
mget /text /as-name:\v(filename).new *.c
This gets all ".c" files and stores them with "
.new" appended to their names. See the [315]C-Kermit 7.0 Update
Notes for details.
/COMMAND
This specifies that the incoming file is to be written to the
standard input of a command, rather than to a file. The command
name is the as-name from the GET command or the /AS-NAME
argument. If you need to refer to the incoming file's name in
the command, use \v(filename). See the description of the
regular Kermit [316]GET /COMMAND command for details and
examples.
/QUIET
Transfers the files quietly; don't put up a file-transfer
display.
/ERROR-ACTION:{QUIT,PROCEED}
This switch affects only MGET. If an error occurs with a
particular file, this tells whether to go on to the next file
(PROCEED) or to stop right away and fail (QUIT). The default is
PROCEED.
The file selection switches are:
/EXCEPT:{pattern} or /EXCEPT:{{pattern}{pattern}{...}}
Exception list for MGET; skip downloading any file whose name
matches any of the given patterns (when using the second format,
up to 64 patterns may be specified). [317]CLICK HERE for syntax
details.
/SMALLER-THAN:number
Download only files whose size is smaller than the given number
of bytes (octets). Requires that the FTP server support the SIZE
or MLSD directive.
/LARGER-THAN:number
Download only files whose size is greater than the given number
of bytes. Requires that the FTP server support the SIZE or MLSD
directive.
/NOBACKUPFILES
During MGET, don't download any files whose names end with
backup suffixes (.~n~ where n is a number).
/NODOTFILES
During MGET, don't download any files whose names begin with
period (.). Equivalent to /EXCEPT:{.*}.
/LISTFILE:local-filename
The given file contains a list of files to GET, one per line.
Filenames in the listfile can contain wildcard characters in the
syntax of the server. There is no limit on the number of lines
in the listfile.
/NAMELIST:local-filename
If this switch is given, then instead of actually retrieving the
selected files, the GET command retrieves a list of the names of
the files that would be retrieved, and places it in the
specified file. The resulting file is an ordinary text file,
with one filename per line, suitable for reading by a person, or
processing by a computer program, including Kermit itself (FOPEN
/ FREAD / FWRITE / FCLOSE), and as /FILELIST: file. If the
filename is omitted or given as "-" (dash, hyphen), the list
goes to the screen. NOTE: if you want a copy of the complete
list sent by the server, use SET FTP DEBUG ON, perform an MGET,
and the temporary file containing the list will be kept rather
than deleted (and Kermit tells you its name).
/UPDATE, /COLLISION:keyword
Explained in [318]Section 3.6.2.
/RECURSIVE
This means to try to download an entire directory tree, rather
than just files from a particular directory. In fact, FTP
protocol does not provide a method to request a recursive
download (unless the server supports MLSD; see [319]Section
3.11), so this works only if the FTP server does it anyway,
without being asked, as some do. In this case, Kermit detects
that names in the returned file list contain directory
separators, and therefore attempts to create the needed
directories as the files arrive. But this can work only if the
server is on the same kind of platform as the client, so the
pathname syntax can be recognized, and also because the server
does not switch between text and binary mode, which would be
vital for cross-platform transfers. Use with caution. Synonym:
/SUBDIRECTORIES.
Even when the server does not provide recursive file lists,
[M]GET /RECURSIVE forces Kermit to replicate any directory
structure implied or expressed by the server's file list. For
example:
get somepath/somefile
Gets the file named somefile from the server's somepath
directory and puts it Kermit's current (or download) directory,
whereas:
get /recursive somepath/somefile
creates the path locally and then puts the file in it. Similarly
for MGET:
mget */data/*
downloads all the files in all the data subdirectories of all
the subdirectories of the server's current directory and stores
them locally in Kermit's current (or download) directory,
whereas:
mget /recursive */data/*
re-creates the server's directory structure locally.
The FTP protocol does not include explicit mechanisms for recursion, so
Kermit builds upon what is available. Although an Internet draft
describes a mechanism ("MLSD") that would allow protocol-driven
recursion, similar to Kermit's File Attribute packets (circa 1984), it
has not yet attained RFC or standard status, and servers are not yet
widely available that offer this feature. In the meantime, the
effectiveness of MGET /RECURSIVE depends on the FTP server
implementation. If the server returns a recursive list in response to
the standard NLST command (whose behavior is ill-defined), Kermit's FTP
MGET /RECURSIVE command uses it to re-create the remote directory tree
locally. If the server supports MLSD, C-Kermit 8.0.206 and Kermit 95
2.1 and later are able to sense it automatically and use it, as
described below in [320]Section 3.11.
The /BEFORE:, /AFTER:, /NOT-BEFORE:, and /NOT-AFTER: switches are not
available for downloading because of the confusion with timezones.
Would the given times be in the local timezone, the server's timezone,
or GMT? The FTP server's directory listings show its own local times
but since we don't know what timezone the server is in, there's no way
to reconcile our local times with the server's. Similarly, /PERMISSIONS
can't be preserved in downloads because FTP protocol provides no means
of querying the server for a file's permission.
Source-file disposition switches:
/DELETE
Each file that is downloaded successfully is to be deleted from
the server. Requires the appropriate file access rights on the
server.
/SERVER-RENAME-TO:template
Asks the server to rename each (remote) source file immediately
after, and only if, it is sent correctly. See [321]PUT
/SERVER-RENAME-TO: for details.
Destination-file disposition switches:
/TO-SCREEN
Displays the incoming file on the screen rather than storing it
on disk. If this switch is given, the /RENAME-TO and /MOVE-TO
switches are ignored, the file-transfer display is suppressed,
and the given file(s) is/are shown on the screen. Can be used
with /FILTER, e.g.
get /text /to-screen /filter:more oofa.txt
In fact, you should always use /TO-SCREEN with /FILTER or
/COMMAND when the command would result in displaying the
incoming file on the screen; otherwise C-Kermit would have no
way of knowing to suppress its file transfer display (since it
can't be expected to know what the command or filter does).
/RENAME-TO:template
Each file that is downloaded is to be renamed as indicated if
and only if it was received completely and without error. The
template can be literal text or can contain variables that are
evaluated for each file. For MGET, the text must contain
variables; for GET it can be a literal string. The \v(filename)
variable contains the name of the current file, so:
ftp mget /rename-to:\v(filename).ok *
causes each file that is successfully downloaded to have ".ok"
appended to its name. For details see [322]Section 4.1 of the
[323]C-Kermit 7.0 Update Notes.
/MOVE-TO:text
Just like /RENAME-TO:, except the text denotes the name of a
directory to which successfully downloaded files are to be
moved. If the directory does not exist, it is created.
The file transfer display does not show the /MOVE-TO or /RENAME-TO
value, since the incoming file has not yet been moved or renamed.
3.6.2. Filename Collisions
What should happen if an incoming file has the same name as an existing
file in the same directory? By default, Kermit's FILE COLLISION setting
applies: BACKUP, RENAME, UPDATE, DISCARD, etc, as described in
[324]Using C-Kermit. Kermit's default FILE COLLISION setting is BACKUP
(rename the existing file and store the incoming file under its own
name) and therefore this is also the default FTP collision action.
The name under which an incoming file is to be stored is determined as
follows:
* If an as-name was given, the as-name is used. Otherwise:
* If the client and server platforms are alike or [325]FTP FILENAMES
is set to LITERAL (or the /FILENAMES:LITERAL switch was given for
this download), the incoming filename is used literally. Otherwise:
* The incoming filename is converted to a form that is friendly to
the local platform. For UNIX, for example, incoming filenames that
are all uppercase (as they might be from, say, VMS or an IBM
mainframe) are converted to lowercase.
If the resulting name coincides with the name of a local file that
already exists, we have a filename collision. Collisions are handled
according to the currently selected collision action:
SET FTP COLLISION { BACKUP, RENAME, UPDATE, DISCARD, APPEND, OVERWRITE
}
This establishes a filename collision for FTP, separate from the
Kermit one. The initial FTP collision setting is inherited from
Kermit's FILE COLLISION setting when the first FTP command is
given, but subsequent changes to Kermit's FILE COLLISION setting
do not affect the FTP COLLISION setting. SHOW FTP tells the
current FTP COLLISION setting.
FTP GET /COLLISION:{BACKUP,RENAME,UPDATE,DISCARD,APPEND,OVERWRITE}
Overrides the current FTP COLLISION action for this download
only.
FTP GET /UPDATE
This is equivalent to GET /COLLISION:UPDATE, and is included for
symmetry with PUT /UPDATE
FTP GET /UPDATE and /COLLISION:UPDATE mean to download only those files
whose modification dates on the server are later than those on the
client. Date-time comparisons are done in Coordinated Universal Time
(UTC, GMT, ZULU). The command:
FTP MGET /COLLISION:APPEND /AS-NAME:newfilename *.*
Downloads all matching remote files into a single local file (in
whatever order the server sends them).
3.6.3. Recovery
Recovery is available for downloads too, but there are some differences
from the uploading case described in [326]Section 3.5.3:
* The transfer must be in BINARY mode. It can not be in text mode,
even if the FTP server is on the same kind of platform as Kermit,
and even if there is no character-set translation. The original
download must also have been in binary mode.
* The FTP server must support the REST ("restart") directive.
Unfortunately, this is not a standard command; at this writing, it
is described only in an Internet Draft, not an RFC or Internet
Standard, but nevertheless it is found in several popular FTP
servers, such as [327]ProFTPD.
Here's how download recovery works:
* Kermit checks for conflicting switches, such as /UPDATE, /COMMAND,
or /FILTER. If /RECOVER is given with these switches an error
occurs.
* The prevailing transfer mode (SET FTP TYPE) must be BINARY. If it
is not, the /BINARY switch must have been included with the FTP
[M]GET command.
If the /RECOVER switch is accepted, then for each selected file:
* A SIZE command is sent for the file (using its remote name). If the
reply indicates the file was not found, or the SIZE command was not
understood, or any other kind of error, recovery is canceled (i.e.
the entire file is downloaded).
* If the sizes of the two files are identical, the file is not sent.
Otherwise:
* Kermit sends the REST directive to the server, indicating the size
of the local file. If the server responds affirmatively, Kermit
opens the local file in append mode and appends the incoming data
to it. Otherwise, recovery is canceled and the entire file is
downloaded.
The /RECOVER switch can be included with any FTP GET or MGET command,
even if it specifies a group of files. This lets you resume an
interrupted batch transfer from where it left off. The files that were
already completely sent are skipped, the file that was interrupted is
recovered, and the remaining files are uploaded. BUT... unlike with
uploading, where this can be done with any mixture of text and binary
files, when downloading, it can only be done if all the files are
binary.
It doesn't matter how the original partial file was downloaded -- FTP,
Kermit, HTTP, Zmodem, etc: as long as the preconditions are met, it can
be recovered with FTP [M]GET /RECOVER, or for that matter also with GET
/RECOVER (using Kermit protocol).
[ [328]Top ] [ [329]FTP Top ] [ [330]C-Kermit Home ] [ [331]Kermit Home
]
3.7. Translating Character Sets
A possibly unique feature of Kermit's FTP client is its ability to
convert character sets when transferring files in text mode,
independent of the capabilities of the FTP server, as well as to
translate the character sets of filenames regardless of transfer mode.
For compatibility with existing FTP clients, and because there is a
certain performance penalty, Kermit won't do this unless you ask for
it. If you enable this feature, you need to inform Kermit of the
character set (to be) used on the server and in some cases (explained
below) also the local file character set. This discussion assumes you
know a bit about character sets (as you must if you have to use them);
see Chapter 16 of [332]Using C-Kermit for a detailed treatment. The
Kermit commands for FTP character-set conversion are:
SET FTP CHARACTER-SET-TRANSLATION { ON, OFF }
Whether to translate character sets when transferring text files
with FTP. OFF by default. Set this to ON to enable character-set
translation for subsequent FTP uploads and downloads.
SET FTP SERVER-CHARACTER-SET [333]name
Text character set (to be) used by the server. Most FTP servers
are ignorant of character sets, so all translations are done
unilaterally by Kermit's FTP client. This means that when
downloading files, you must know in advance the character-set
used in the files you are downloading (and in their names). When
uploading, you must specify the character-set to which local
filenames and text-file contents are to be translated for
transmission to the server. If you SET FTP
CHARACTER-SET-TRANSLATION ON but do not specify an FTP
SERVER-CHARACTER-SET, [334]UTF8 is used, since this is the new
Internet standard international character set; it is upwards
compatible with ASCII and it encompasses most written languages
and therefore does not favor any particular group of people, as
any other default would do. If you SET FTP SERVER-CHARACTER-SET
to something (anything) when FTP CHARACTER-SET TRANSLATION is
OFF, this also sets the latter ON.
SET FILE CHARACTER-SET [335]name
This is the regular Kermit (non-FTP-specific) command for
identifying the character set (to be) used in local text files
and filenames.
TO REITERATE: If you SET FTP CHARACTER-SET TRANSLATION ON but do not
specify an FTP SERVER-CHARACTER-SET, outbound text files are converted
to UTF-8 and inbound text files are assumed to be UTF-8. If this is not
appropriate, be sure to also specify the desired FTP
SERVER-CHARACTER-SET.
You can use "special" (non-ASCII) characters in filenames in all the
client / server file management commands (FTP MKDIR, RMDIR, DIRECTORY,
VDIRECTORY, DELETE, etc), and also in file-transfer commands. When
giving commands such as FTP DIR (RDIR) and FTP PWD (RPWD), the reply is
translated too, so you can read it. In this example, the client and
server use entirely different codes to represent the special characters
of German:
C-Kermit> ftp xyzcorp.de /anonymous
C-Kermit> set ftp server-character-set latin1
C-Kermit> set file character-set german
C-Kermit> rcd Städte
C-Kermit> rpwd
"/pub/ftp/Städte is current directory"
C-Kermit> rdir
-rw-rw---- 1 olaf 54018 Jan 6 17:58 Adenbüttel.txt
-rw-rw---- 1 ursula 373 Jan 5 15:19 Aßlar.txt
-rw-rw---- 1 gisbert 482 Jan 5 15:20 Blowatz.txt
-rw-rw---- 1 gudrun 124 Jan 5 15:19 Böblingen.txt
-rw-rw---- 1 olga 14348 Jan 7 14:23 Köln.txt
When the client and server file systems use different character sets,
you should take care to use only those characters that the two sets
share in common when creating filenames or text-file contents. For
example, PC code pages contain a lot line- and box-drawing characters,
and sometimes "smart quotes", etc, that are not found in ISO standard
8-bit character sets. You should be especially careful to avoid using
such characters in filenames.
[ [336]C-Kermit Character Sets ]
3.7.1. Character Sets and Uploading
Kermit's PUT and MPUT commands include full file-scanning capabilities,
as described in [337]Section 4. Thus if FTP CHARACTER-SET-TRANSLATION
is ON and your character-set associations are set up appropriately,
Kermit automatically switches on a per-file basis between text and
binary mode, and for each text file between your chosen 7-bit text
character set (e.g. ASCII or ISO 646 German), 8-bit text (e.g. Latin-1
or Japanese EUC), UCS-2, and UTF-8, and converts each of these
automatically to the server character-set, and furthermore
automatically differentiates between the Little and Big Endian forms of
UCS-2, always sending in Big Endian form.
WARNING: It is not advisable to use UCS-2 (or any Unicode
transformation other than UTF-8) "on the wire", i.e. as a server
character set. Most FTP servers are not able to cope with it, since
it contains lots of 0 (NUL) characters. If you do use it, Kermit
does not translate filenames to or from UCS-2, for reasons well
known to C programmers (for example, UNIX APIs assume filename
strings are NUL-terminated). [338]UTF-8 is the preferred (and
standard) Unicode format for the Internet.
FTP character-set translations differ from the regular Kermit ones by
not restricting translations to a file-character-set /
transfer-character-set pair. You can have Kermit's FTP client translate
between any pair of character sets it knows about. You can see the list
of supported character sets by typing either of the following:
set ftp server-character-set ?
set file character-set ?
A typical list looks like this ([339]CLICK HERE for an explanation of
the names):
C-Kermit>set file char ? One of the following:
ascii cp869-greek hebrew-7 mazovia-pc
british cyrillic-iso hebrew-iso next-multinational
bulgaria-pc danish hp-roman8 norwegian
canadian-french dec-kanji hungarian portuguese
cp1250 dec-multinational iso2022jp-kanji shift-jis-kanji
cp1251-cyrillic dg-international italian short-koi
cp1252 dutch jis7-kanji spanish
cp437 elot927-greek koi8 swedish
cp850 elot928-greek koi8r swiss
cp852 euc-jp koi8u ucs2
cp855-cyrillic finnish latin1-iso utf8
cp858 french latin2-iso
cp862-hebrew german latin9-iso
cp866-cyrillic greek-iso macintosh-latin
C-Kermit>
Thus you can translate not only between private sets (like PC code
pages) and standard ones (like Latin-1) as in Kermit protocol, but also
between any given pair of private sets (e.g. CP852 and Mazovia). All
conversions go through Unicode as the intermediate character set,
resulting in a minimum of character loss, since Unicode is a superset
of all other character sets known to Kermit.
In addition to the SET commands listed above, the FTP PUT and MPUT
commands include switches that apply only to the current command:
/LOCAL-CHARACTER-SET:name
/SERVER-CHARACTER-SET:name
Use these switches to force a particular translation. These
switches override the global FTP CHARACTER-SET-TRANSLATION and
SERVER-CHARACTER-SET settings and also character-set
differentiation by file scanning for the duration of the PUT or
MPUT command. The file scan is still performed, however, to
determine whether the file is text or binary; thus these
switches do not affect binary files unless you also include the
/TEXT switch to force all files to be treated as text.
In other words, if you include one or both of these switches with a PUT
or MPUT command, they are used. Similarly, the /TRANSPARENT switch
disables character-set translation for the PUT or MPUT command despite
the prevailing FTP CHARACTER-SET-TRANSLATION and SERVER-CHARACTER-SET
settings.
When uploading, the FILE CHARACTER-SET setting is ignored unless you
have forced Kermit not to [340]scan local files by including a /TEXT or
/BINARY switch with your [M]PUT command, or by disabling automatic
text/binary switching in some other way.
Examples:
1. Suppose you have a CP852 (East European) text file that you want to
upload and store in ISO Latin Alphabet 2 encoding:
ftp put /local-char:cp852 /server-char:latin2 magyar.txt
2. Suppose you always want your text files converted to Latin-2 when
uploading with FTP. Then put:
set ftp server-character-set latin2
in your Kermit customization file, and then you can omit the
/SERVER-CHARACTER-SET: switch from your FTP PUT commands:
ftp put /local-char:cp852 magyar.txt
3. Now suppose that all the text files on your PC are written in
Hungarian, but they have a variety of encodings, and you don't want
to have to include the /LOCAL-CHARACTER-SET: switch on every FTP
PUT command, or (more to the point) you want to be able to send a
mixture of these files all at once. Put these commands in your
Kermit customization file:
set ftp server-character-set latin2 ; ISO 8859-2
set file default 7-bit-character-set hungarian ; ISO 646 Hungarian
set file default 8-bit-character-set cp852 ; PC East European Code Page
and now PUT and MPUT will automatically detect and switch among ISO
646 Hungarian, Code Page 852, UTF-8, and UCS-2 encodings,
translating each one to Latin-2 for uploading:
ftp put *.txt
And since binary files are also detected automatically, the latter can
be simplified to:
ftp put *
even when "*" matches a diverse collection of binary and text files,
because translations are skipped automatically for binary files.
3.7.2. Character Sets and Downloading
The commands and switches are the same as for uploading, but automatic
character-set switching works differently, since Kermit can't scan the
server files in advance. Instead, the transfer mode (text or binary) is
based on the filenames; each name is compared with Kermit's list of
text name patterns and binary name patterns. If the name matches a
binary pattern (for example, if the filename is oofa.tar.gz and one of
the filename patterns is "*.gz"), the file is downloaded in binary
mode; otherwise if it matches a text pattern (e.g. oofa.txt matches
"*.txt"), it is transferred in text ("ascii") mode. Otherwise, it is
transferred in the prevailing FTP TYPE.
In C-Kermit 8.0, the pattern lists used with FTP GET are not the same
lists used with Kermit transfers, and can not be viewed with SHOW
PATTERNS, nor adjusted with ADD and REMOVE TEXT-PATTERNS and
BINARY-PATTERNS, or SET FILE TEXT-PATTERNS and BINARY-PATTERNS.
Configuration of the FTP patterns list will be added in a future
release.
Examples:
get /server-char:latin1 /local-char:cp850 Grüße.txt
In this command, the filename contains special characters, which
you enter using whatever character set your local computer uses,
in this case PC Code Page 850 (cp850). The command tells Kermit
(in case it didn't know already from its FILE CHARACTER-SET
setting) that the local character set is cp850 and the server's
character-set is ISO 8859-1 Latin Alphabet 1 (latin1). Kermit
translates the filename from cp850 to latin1 and sends the
latin1 name to the server. Since it's a text file (matches
"*.txt"), its contents are translated to cp850 on arrival, and
it is saved with a cp850 name.
mget /text /server:latin1 /local:utf8 *.txt
This command:
+ Tells C-Kermit that the server's files are encoded in ISO
8859-1 Latin Alphabet 1.
+ Tells C-Kermit to translate the incoming files into Unicode
UTF-8 for storage.
+ Asks the server to send all ".txt" files in text mode.
mget /server:latin1 /local:utf8 *
Tells Kermit to get all files from the server's directory,
switching between text and binary mode based on the filename.
The names of all the files are translated (to UTF-8 in this
case), but contents are translated (also to UTF-8) only for text
files.
Note that any pair of 8-bit character sets is likely to have some
incompatibilities. Any characters in the source file that do not have
equivalents in the destination file's character set are converted to
question marks. This applies to both filenames and to text file
contents.
Also note that the server's ability to accept special characters in
filenames depends on the particular server. For example:
get Grüße.txt
works with WU-FTPD, but:
mget Grüß*.txt
does not.
3.7.3. RFC2640
[341]RFC2640, July 1999, specifies a method by which the FTP client and
server can negotiate the use of UTF8. However, RFC2640-capable servers
are rare to nonexistent at this writing, and in any case you don't need
them to be able to transfer text in UTF8. C-Kermit lets you upload and
download text files in any character set it knows about, converting to
or from any other character set it knows about, without the knowledge,
permission, or cooperation of the server, and regardless of its
capabilities.
[ [342]Top ] [ [343]FTP Top ] [ [344]C-Kermit Home ] [ [345]Kermit Home
]
3.8. FTP Command Shortcuts
C-Kermit's FTP client coexists with other C-Kermit functions by
requiring the "ftp" prefix for each FTP-related command: FTP OPEN, FTP
GET, FTP BYE, and so on. For interactive use, however, this can be
rather awkward and sometimes surprising, for example when a GET command
starts a Kermit GET rather than an FTP GET. In fact, many Kermit
commands might just as easily apply to an FTP connection: GET, PUT
(SEND), BYE, and CLOSE. The following command lets you choose how these
commands are interpreted:
SET GET-PUT-REMOTE { AUTO, KERMIT, FTP }
Controls the orientation of GET, PUT, REMOTE and other
file-transfer and client/server commands that might apply to
either Kermit or FTP. The default setting is AUTO, meaning that
these commands apply to FTP if an FTP connection is open, and to
Kermit otherwise. KERMIT means they always apply to Kermit, FTP
means they always apply to FTP.
Here is a complete list of affected commands:
Kermit Command FTP Equivalent
(none) FTP [ OPEN ]
LOGIN FTP USER
LOGOUT FTP RESET
BYE FTP BYE
FINISH FTP BYE
CLOSE FTP BYE
HANGUP FTP BYE
BINARY FTP TYPE BINARY
TEXT (or ASCII) FTP TYPE ASCII
SEND (or PUT) FTP PUT
MSEND (or MPUT) FTP MPUT
RESEND FTP PUT /RECOVER
CSEND FTP PUT /COMMAND
GET FTP GET
MGET FTP MGET
REGET FTP GET /RECOVER
REMOTE HELP (RHELP) FTP HELP
REMOTE CD (RCD) FTP CD (CWD)
REMOTE PWD (RPWD) FTP PWD
REMOTE DIRECTORY (RDIR) FTP DIRECTORY
REMOTE DELETE (RDEL) FTP DELETE
REMOTE MKDIR (RMKDIR) FTP MKDIR
REMOTE RMDIR (RRMDIR) FTP RMDIR
REMOTE RENAME (RRENAME) FTP RENAME
REMOTE TYPE (RTYPE) FTP TYPE
REMOTE EXIT (REXIT) FTP BYE
The commands in the right-hand column always access FTP. The commands
in the left column can access either Kermit protocol or FTP:
* When GET-PUT-REMOTE is set to KERMIT, or to AUTO when there is no
FTP connection, the commands in the left-hand column access Kermit
protocol, and those right-hand column are required for FTP.
* When GET-PUT-REMOTE is set to FTP, or to AUTO when there is an
active FTP connection, the commands in the left-hand column access
the FTP connection and can not be used to access Kermit protocol.
In this case, if you want to be able to use both Kermit protocol
and the FTP connection, you must SET GET-PUT-REMOTE KERMIT, and
then use the FTP commands in the right-hand column to access the
FTP connection.
Note that file-management commands such as DIRECTORY, DELETE, CD, PWD,
MKDIR, RMDIR, HELP, RENAME, COPY, TYPE, and so on, always apply
locally, no matter what kind of connection you have. This is the
opposite of most FTP clients, where these commands are intended for the
server, and require an "L" prefix for local execution (e.g. "dir" gets
a directory listing from the server, "ldir" gets a local directory
listing). To illustrate with the CD command and a typical UNIX FTP
client:
Client Server Change Local Directory Change Remote Directory
FTP FTP lcd cd (cwd)
Kermit Kermit cd rcd, remote cd
Kermit FTP cd ftp cd, rcd, remote cd
Also note that not all REMOTE commands are useful with FTP, since FTP
servers do not offer the corresponding functions. These include:
* REMOTE ASSIGN - FTP servers don't have variables
* REMOTE COPY - FTP servers don't copy files
* REMOTE HOST - FTP servers don't execute host (shell) commands
* REMOTE KERMIT - FTP servers don't execute Kermit commands
* REMOTE PRINT - FTP servers don't print files
* REMOTE QUERY - FTP servers don't have variables
* REMOTE SET - FTP servers don't have Kermit settings
* REMOTE WHO - FTP servers don't send user lists
Finally note that command shortcuts do not apply to the HELP command.
For help about an FTP command, use (for example) "help ftp delete", not
"help delete" or "help rdelete".
[ [346]Top ] [ [347]FTP Top ] [ [348]C-Kermit Home ] [ [349]Kermit Home
]
3.9. Dual Sessions
You can have an FTP session open at the same time as a regular Kermit
SET LINE or SET HOST (terminal) session. In this case, the default SET
GET-PUT-REMOTE AUTO setting should ensure that all "two-faced" commands
like GET, PUT, REMOTE, HANGUP, BYE, etc, apply to the Kermit session,
and all commands for the FTP session must include the FTP prefix. To be
absolutely certain, you can use SET GET-PUT-REMOTE KERMIT.
ftp foo.bar.baz.com
if fail ...
(log in)
set host foo.bar.baz.com
if fail ...
(log in)
Now you have both an FTP and Telnet connection to the same host (of
course they could also be to different hosts, and you could also have a
direct or dialed serial connection instead of a Telnet connection). Now
assuming you have a Kermit server on the far end of the Kermit
connection:
rcd incoming ; Changes Kermit server's directory (= REMOTE CD)
ftp cd incoming ; Changes FTP server's directory
put oofa.txt ; Sends a file on the Kermit connection
ftp put oofa.txt ; Sends a file on the FTP connection
bye ; Shuts down the Kermit connection
ftp bye ; Shuts down the FTP connection
Note that PUT and SEND are synonyms for both FTP and Kermit
connections.
You can also establish dual sessions on the Kermit command line:
kermit -j host1 -9 host2
This makes a Telnet connection to host1 and an FTP connection to host2.
[ [350]Top ] [ [351]FTP Top ] [ [352]C-Kermit Home ] [ [353]Kermit Home
]
3.10. Automating FTP Sessions
Most of Kermit's scripting features can be used to make and control FTP
sessions: FOR and WHILE loops, IF-ELSE and SWITCH constructions,
variables, arrays, built-in functions, and all the rest. You can't use
INPUT, MINPUT, OUTPUT, CLEAR, or SCRIPT on an FTP session, but these
are not needed since the FTP protocol is well defined.
[354]CLICK HERE for an FTP scripting tutorial.
3.10.1. FTP-Specific Variables and Functions
The following variable tells whether an FTP connection is open:
\v(ftp_connected)
1 if there is an active FTP connection, 0 if there isn't.
The FTP OPEN command sets:
\v(ftp_host)
The host to which the most recent FTP connection was made.
\v(ftp_security)
The security method negotiated for the current FTP session. The
value is "NULL" when no security is used. Other possibilities
are GSSAPI, KERBEROS_V4, SSL, TLS, and SRP. Also see
\v(authname), \v(authstate), and \v(authtype). See [355]3.2.
Making Secure FTP Connections.
\v(ftp_server)
The OS type (UNIX, VMS, etc) of the FTP server host.
The FTP USER command (or FTP OPEN /USER:, or FTP with automatic login)
sets:
\v(ftp_loggedin)
1 if you are logged in to an FTP server, 0 if you are not.
The current COMMAND-PROTECTION-LEVEL and DATA-PROTECTION-LEVEL values
are reflected in:
\v(ftp_cpl)
\v(ftp_dpl)
The values are "clear", "confidential", "safe" or "private". See
[356]3.2. Making Secure FTP Connections.
The FTP GET-PUT-REMOTE setting is reflected in:
\v(ftp_getputremote)
The values are "auto", "ftp", or "kermit".
Every FTP command sets the \v(success) variable, as well as the
following two FTP-specific variables:
\v(ftp_code)
The standardized numeric FTP protocol code from the server's
response to the last client command, a 3-digit decimal number
defined in [357]RFC959. Briefly:
1xx = Positive Preliminary Reply
2xx = Positive Completion Reply
3xx = Positive Intermediate Reply
4xx = Transient Negative Completion Reply
5xx = Permanent Negative Completion Reply
\v(ftp_message)
The text message, if any, from the server's response to the last
client command. If the most recent response had multiple lines,
this variable has only the final line. These messages are not
standardized and vary in format and content from server to
server. Synonym: \v(ftp_msg).
FTP file transfers set the regular Kermit transfer status variables:
\v(cps) Characters per second of most recent transfer.
\v(filespec) File specification used in most recent transfer.
\v(fsize) Size of file most recently transferred.
\v(tfsize) Total size of file group most recently transferred.
\v(xferstatus) Status of most recent transfer (0 = success, 1 = failure).
\v(tftime) Elapsed time of most recent transfer, in seconds.
During an FTP transfer, the per-file variables are:
\v(filename) Name of current file.
\v(filenumber) Ordinal file number in group (1, 2, 3, ...)
3.10.2. Examples
Let's begin with a simple example showing how to log in, send some
files, and log out:
define error if fail { ftp bye, stop 1 Error: \%1 }
set transact brief
log t
ftp ftp.xyzcorp.com /anonymous
if fail stop 1 Connection failed
if not \v(ftp_loggedin) stop 1 Login failed
ftp cd incoming
error {ftp cd}
cd upload
error {local cd}
ftp put /delete *
error {put}
ftp bye
First we define an error handling macro to be used after the connection
is made. Then we set up a brief-format transaction log to keep a record
of our file transfers. Then we make a connection to the host and log in
anonymously. The "if fail" command checks whether the connection was
made. The "if not" command checks whether login was successful.
Obviously the script should not continue unless both tests succeed.
Next we change to the server's 'incoming' directory and to our own
'upload' directory, and send all the files that are in it (they can be
any mixture of text and binary files), deleting each source file
automatically after it is successfully uploaded. Each of these
operations is checked with the ERROR macro, which prevents the script
from continuing past a failure.
Finally we close the FTP session with the "bye" command.
Just like any other Kermit script, this one can be used in many ways:
* It can be stored in a file, and Kermit can be told to TAKE the
file.
* In UNIX, it can be a "[358]kerbang" script and therefore run
directly from the shell prompt or as a cron job.
We could have used command shortcuts like "rcd", "put", and "bye", but
since they can be ambiguous under certain circumstances, it is better
to avoid them in scripts; they are intended mainly for convenience
during interactive use. However, if you wish to use the shortcuts in a
script, you can do it this way (error handling omitted for brevity):
local \%t ; Declare a local temporary variable
assign \%t \v(ftp_getputremote) ; Save current FTP GET-PUT-REMOTE setting
set ftp get-put-remote ftp ; Choose FTP orientation
ftp xyzcorp.com /anonymous ; Open an FTP connection
get oofa.txt ; GET a file
put foo.bar ; PUT a file
rdel yesterday.log ; Delete a file on the server
bye ; Log out and disconnect from server.
set ftp get-put-remote \%t ; Restore previous GET-PUT-REMOTE setting
Of course, FTP scripts can also be written as macros. This lets you
pass parameters such as hostnames, usernames, and filenames to them:
define doftpget {
if < \v(argc) 4 end 1 Usage: \%0 host user remotefile [ localfile ]
ftp \%1 /user:\%2
if fail end 1 FTP OPEN \%1 failed
if not \v(ftp_loggedin) end 1 FTP LOGIN failed
ftp get {\%3} {\%4}
if fail end 1 FTP GET \%3 failed
ftp bye
}
Add this definition to your Kermit customization file, and it will
always be available when you start Kermit. This macro lets you download
a file with FTP by giving a single command, e.g.:
doftpget xyzcorp.com anonymous oofa.txt
3.10.3. Automating Secure FTP Sessions
Often when making secure connections, you are prompted interactively
for certain information or permission to proceed. These prompts can
stop an automated procedure. To avoid them, you must give the
appropriate commands to disable them, and/or supply the prompted-for
information beforehand. Here are a few hints:
* Make sure that SET TAKE ERROR and SET MACRO ERROR are both OFF.
This is the default, but in case you have set either one of these
ON in your script or initialization file, this makes the script
halt on any kind of error. Normally you would want to check each
operation for success or failure and take appropriate action.
* On SSL and TLS connections, you may be asked whether it is OK to
proceed with a connection to server that presents a self-signed
certificate. You can use the SET AUTHENTICATION SSL (or TLS) VERIFY
or SET AUTH SSL (or TLS) CERTS-OK commands to avoid this prompt by
not requesting a certificate from the peer.
* (More to be added...)
[ [359]Top ] [ [360]FTP Top ] [ [361]FTP Script Tutorial ] [
[362]C-Kermit Home ] [ [363]Kermit Home ]
3.11. Advanced FTP Protocol Features
The remainder of the FTP documentation (through the end of Section 3)
is new to C-Kermit 8.0.206, but we leave it in black to prevent
headaches. Except for titles.
* [364]TERMINOLOGY
* [365]FEATURE NEGOTIATION
* [366]USING MGET: NLST VERSUS MLSD
* [367]EXAMPLES
* [368]REFERENCES
The new releases of [369]C-Kermit (8.0.206) and [370]Kermit 95 (2.1)
support new FTP protocol features from RFC 2389 as well as most of
what's in the Elz and Hethmon Extensions to FTP Internet Draft (see
[371]References). Some of these features, such as SIZE (request a
file's size), MDTM (request file's modification time), and REST
(restart interrupted transfer) have been widely implemented in FTP
clients and servers for years (as well as in the initial release of the
Kermit FTP clients). Others such as FEAT and MLSD are rarely seen and
are new to the upcoming Kermit releases. TVFS (Trivial Virtual File
Store) is supported implicitly, and the UTF-8 character-set is already
fully supported at the protocol and data-interchange level.
For Kermit users, the main benefit of the new FTP protocol extensions
is the ability to do recursive downloads. But the extensions also
introduce complications and tradeoffs that you should be aware of. Of
course Kermit tries to "do the right thing" automatically in every case
for backwards compatibility. But (as noted later) some cases are
inherently ambiguous and/or can result in nasty surprises, and for
those situations new commands and switches are available to give you
precise control over Kermit's behavior, in case the defaults don't
produce the desired results.
3.11.1. Terminology
Command-line FTP clients such as Kermit (as well as the traditional FTP
programs found on Unix, VMS, ..., even Windows) have commands like PUT,
MPUT, GET, MGET, and BYE, which they convert into zero or more FTP
protocol commands, such as NLST, RETR, QUIT. For clarity, we'll use
"command" to refer to commands given by the user to the FTP client, and
"directive" for FTP protocol commands sent by the FTP client to the FTP
server.
3.11.2. Feature Negotiation
New FTP protocol features are negotiated by the client sending a FEAT
directive and the server responding with a list of (new) features it
supports, or else with an error indication if it does not support the
FEAT directive at all, in which case the client has to guess which new
features it supports (Kermit guesses that it supports SIZE and MDTM but
not MLST). Note that the MLST feature includes MLSD, which is not
listed separately as a feature.
Guessing is nice when it works, but sometimes it doesn't, and some FTP
servers become confused when you send them a directive they don't
understand, or they do something you didn't want, sometimes to the
point of closing the connection. For this reason, Kermit lets you
override default or negotiated features with the following new
commands:
FTP { ENABLE, DISABLE } FEAT
Enables or disables the automatic sending of a FEAT directive
upon connection to an FTP server. Note that FTP [ OPEN ] /NOINIT
also inhibits sending the FEAT directive (and several others)
for the connection being OPEN'd, but without necessarily
disabling FEAT for subsequent connections in the same Kermit
instance. FEAT is ENABLED by default, in which case many FTP
servers are likely to reply:
500 'FEAT': command not understood
which is normally harmless (but you never know). (In C-Kermit
8.0.208, this error message is suppressed unless you SET FTP
DEBUG ON.)
FTP ENABLE { MDTM, MLST, SIZE }
Enables the given directive for implicit use by the FTP GET and
MGET commands in case it has been disabled or erroneously
omitted by the server in its FEAT response. Note: MLSD can be
used in the FTP ENABLE and DISABLE commands as a synonym for
MLST. YOU MUST GIVE THIS COMMAND AFTER MAKING THE FTP
CONNECTION.
FTP DISABLE { MDTM, MLST, SIZE }
Disables implicit use of the given directive by GET or MGET in
case it causes problems; for example, because it makes multifile
downloads take too long or the server announces it erroneously
or misimplements it. Use DISABLE FEAT before making a connection
to prevent Kermit from sending the FEAT directive as part of its
initial sequence. Note that disabling FEAT, SIZE, or MDTM does
not prevent you from executing explicit FTP FEATURES, FTP SIZE,
or FTP MODTIME commands. Also note that disabling SIZE prevents
PUT /RESTART (recovery of interrupted uploads) from working. YOU
MUST GIVE THIS COMMAND AFTER MAKING THE FTP CONNECTION.
To enable or disable more than one feature, use multiple FTP ENABLE or
FTP DISABLE commands. The SHOW FTP command shows which features are
currently enabled and disabled.
FTP FEATURES
This command sends a FEAT directive to the server. In case you
have been disabling and enabling different features, this
resynchronizes Kermit's feature list with the server's. If the
server does not support the FEAT directive, Kermit's feature
list is not changed.
FTP OPTIONS directive
Informational only: the server tells what options, if any, it
supports for the given directive, e.g. MLST. Fails if the server
does not support the OPTS directive or if the directive for
which options are requested is not valid. The directive is
case-insensitive.
FTP SIZE filename
Sends a SIZE directive to the server for the given file. The
filename must not contain wildcards. The server responds with an
error if the file can't be found, is not accessible, or the SIZE
directive is not supported, otherwise with the length of the
file in bytes, which Kermit displays and also makes available to
you in its \v(ftp_message) variable. If the directive is
successful, Kermit (re-)enables it for internal use by the GET
and MGET directives on this connection.
FTP MODTIME filename
Works just like the SIZE directive except it sends an MDTM
directive. Upon success, the server sends modification date-time
string, which Kermit interprets for you and also makes available
in its \v(ftp_message) variable.
Whenever a SIZE or MDTM directive is sent implicitly and rejected by
the server because it is unknown, Kermit automatically disables it.
3.11.3. Using MGET: NLST versus MLSD
When you give an MGET command to an FTP client, it sends a request to
the FTP server for a list of files, and then upon successful receipt of
the list, goes through it and issues a RETR (retrieve) directive for
each file on the list (or possibly only for selected files).
With the new FTP protocol extensions, now there are two ways to get the
list of files: the NLST directive, which has been part of FTP protocol
since the beginning, and the new MLSD directive, which is new and not
yet widely implemented. When NLST is used and you give a command like
"mget *.txt", the FTP client sends:
NLST *.txt
and the server sends back a list of the files whose names match, e.g.
foo.txt
bar.txt
baz.txt
Then when downloading each file, the client sends SIZE (if it wants
have a percent-done display) and MDTM (if it wants to set the
downloaded file's timestamp to match that of the original), as well as
RETR (to retrieve the file).
But when MLSD is used, the client is not supposed to send the filename
or wildcard to the server; instead it sends an MLSD directive with no
argument (or the name of a directory), and the server sends back a list
of all the files in the current or given directory; then the client
goes through the list and checks each file to see if it matches the
given pattern, the rationale being that the user knows only the local
conventions for wildcards and not necessarily the server's conventions.
So with NLST the server interprets wildcards; with MLSD the client
does.
The interpretation of NLST wildcards by the server is not
necessarily required or even envisioned by the FTP protocol
definition (RFC 959), but in practice most clients and servers work
this way.
The principal advantage of MLSD is that instead of sending back a
simple list of filenames, it sends back a kind of database in which
each entry contains a filename together with information about the
file: type, size, timestamp, and so on; for example:
size=0;type=dir;perm=el;modify=20020409191530; bin
size=3919312;type=file;perm=r;modify=20000310140400; bar.txt
size=6686176;type=file;perm=r;modify=20001215181000; baz.txt
size=3820092;type=file;perm=r;modify=20000310140300; foo.txt
size=27439;type=file;perm=r;modify=20020923151312; foo.zip
(etc etc...)
(If the format of the file list were the only difference between NLST
and MLSD, the discussion would be finished: it would always be better
to use MLSD when available, and the MGET user interface would need no
changes. But there's a lot more to MLSD than the file-list format; read
on...)
The client learns whether the server supports MLSD in FEAT exchange.
But the fact that the server supports MLSD doesn't mean the client
should always use it. It is better to use MLSD:
* On connections where the server imposes a time penalty for every
command, e.g. the Red Hat Rawhide server. With MLSD, the client
needs to send only one command (RETR) per file, whereas NLST
requires three (SIZE, RETR, and MDTM). Suppose there is a 30-second
delay for each command and 1000 files are to be fetched; in that
case, MLSD saves 60,000 seconds = 1000 minutes = 16 hours and 40
minutes.
* For recursive downloads since there is no dependable way to
download directory trees with NLST.
But it is better to use NLST:
* If you want only a couple short files out of a large directory. In
this case, NLST is the better choice since the server sends a list
of only the files you want, not a list of (say) a million files,
which can make a big difference on slow connections. For example,
suppose your wildcard matches three files of 1K each, but the
million-file listing is 80MB long, and your connection is through a
modem. The overhead of using MLSD is practically infinite.
* If the server supports wildcarding features not known to the
client, but that can be used to achieve desirable effects otherwise
unobtainable, such as "[dir...]*.txt" in VMS or AOS/VS "except"
clauses.
* If you have been given a wildcard string by an FTP site
administrator for fetching a specific group of files out of a
larger directory, e.g. "mget ck[cuw]*.[cwh] makefile", that is
expected to work with any client (an FTP site administrator can't
be expected to know the wildcard syntax of every FTP client).
But when using MLSD there are complications:
* MLSD wants either a blank argument (meaning the current directory)
or else the name of a specific directory. The client must not send
it a wildcard or a filename.
* But if the user's command is "mget xxx", how does the client know
whether to send "xxx" in the MLSD directive? It might be the name
of a directory on on the server, in which case it should be sent,
or it might be the name of a file on the server (or a wildcard), in
which case it must not be sent. Since the client knows its own
wildcard syntax, then in most cases it would be right to send
"MLSD" with no argument if xxx is wild, and to send "MLSD xxx" if
it is not.
* But suppose the server's file system allows filename characters
that correspond with the client's wildcard syntax? For example:
"[abc]" could be either a valid VMS directory name or a wildcard
pattern used by the FTP client. What should the client do with
"mget [abc]"? In this case there must be a way for the user to
force sending the MGET argument as the MLSD argument.
* If "xxx" is a regular file in the server's current directory, "mget
xxx" works with NLST but not with MLSD.
To further complicate matters, NLST can (in theory) work just like
MLSD: if sent with a blank argument or a directory name, it is supposed
to return a complete list of files in the current or given directory,
which the client can match locally against some pattern. It is not
known if any FTP server or client does this but nevertheless, it should
be possible since this behavior can be inferred from RFC 959.
In view of these considerations, and given the need to preserve the
traditional FTP client command structure and behavior so the software
will be usable by most people:
1. The MGET command should produce the expected result in the common
cases, regardless of whether NLST or MLSD is used underneath.
2. For anomalous cases, the user needs a way to control whether the
MGET argument is sent to the server or kept for local use.
3. At the same time, the user might need a way to send a directory
name to the server, independent of any wildcard pattern.
4. The user needs a way to force NLST or MLSD for a given MGET
command.
By default, Kermit's MGET command uses MLSD if MLST is reported by the
server in its FEAT list. When MLSD is used, the filespec is sent to the
server if it is not wild (according to Kermit's own definition of
"wild" since it can't possibly know the server's definition). If the
filespec is wild it is held for local use to select files from the list
returned by the server. If MLST is not reported by the server or is
disabled, Kermit sends the MGET filespec with the NLST directive.
The default behavior can be overridden globally with FTP DISABLE MLST,
which forces Kermit to use NLST to get file lists. And then for
situations in which MLSD is enabled, the following MGET switches can be
used to override the defaults for a specific MGET operation:
/NLST
Forces the client to send NLST. Example:
mget /nlst foo.*
/MLSD
Forces the client to send MLSD (even if MLST is disabled).
Example:
mget /mlsd foo.*
/MATCH:pattern
When this switch is given, it forces the client to hold the
pattern for local use against the returned file list. If a
remote filespec is also given (e.g. the "blah" in "mget
/match:*.txt blah"), then it is sent as the NLST or MLSD
argument, presumably to specify the directory whose files are to
be listed. When the /MATCH switch is not given, the MGET
filespec is sent to the server if the directive is NLST or if
the filespec is not wild. Examples:
Command: With NLST: With MLSD:
mget NLST MLSD
mget *.txt NLST *.txt MLSD
mget foo NLST foo MLSD foo
mget /match:*.txt NLST MLSD
mget /match:*.txt foo NLST foo MLSD foo
In other words, the pattern is always interpreted locally unless MGET
uses NLST and no /MATCH switch was given.
3.11.4. Examples
3.11.4.1. Downloading a Single File
There are no choices here, just use the FTP GET command. Kermit always
sends the RETR directive, and possibly SIZE and/or MDTM. The small
advantage of using MLST in this case is outweighed by the risk and
effort of coding a special case.
3.11.4.2. Downloading a Group of Files from a Single Directory
This case presents tradeoffs, especially on slow connections:
* For downloading all or most of the files in a directory, MLSD is
better because it eliminates the need to send SIZE and MDTM for
each file. No special actions are required in this case; Kermit
uses MLSD automatically if the server supports it (unless you have
disabled it).
* For a small number of files from a large directory, NLST is better
because it bypasses downloading of a potentially huge file list
prior to the files themselves. If you have a connection to a server
that supports MLSD, use the /NLST switch to force NLST:
mget /nlst t[1234].h
* If the server supports MLSD but does not support separate SIZE or
MDTM directives, and you need the size and/or timestamp
information, MLSD is better; no special actions required.
* If the server supports MLSD but does not support the "size" and
"modify" facts, but it does support the SIZE or MDTM directives,
and you need the size and/or timestamp information, NLST is better.
3.11.4.3. Downloading a Directory Tree
MLSD is the only choice for recursive downloads; they rarely, if ever,
work with NLST (the few cases where they do work rely on extra-protocol
"secret" notations for the NLST argument). No special actions are
required to force MLSD when the server supports it, unless you have
disabled it. Examples:
MGET /RECURSIVE
This tells the server to send all files and directories in the
tree rooted at its current directory.
MGET /RECURSIVE *.txt
This tells the server to send all *.txt files in the tree rooted
at its current directory.
MGET /MLSD /RECURSIVE *.txt
Same as the previous example but forces Kermit to send MLSD in
case it was disabled, or in case the server is known to support
it even though it did not announce it in its FEAT listing.
MGET /RECURSIVE /MATCH:*.zip archives
Tells the server to send all ZIP files in the tree rooted at its
"archives" directory.
MGET /RECURSIVE /MATCH:* [abc]
The server is running on VMS and you want it to send all the
files in the directory tree rooted at [ABC]. But since "[abc]"
looks just like a wildcard, you have to include a /MATCH: switch
to force Kermit to send "[abc]" as the MLSD argument.
In all cases in which the /RECURSIVE switch is included, the server's
tree is duplicated locally.
Although MLSD allows recursion and NLST does not, the MLSD
specification places a heavy burden on the client; the obvious,
straightforward, and elegant implementation (depth-first, the one
that Kermit currently uses) requires as many open temporary files as
the server's directory tree is deep, and therefore client resource
exhaustion -- e.g. exceeding the maximum number of open files -- is
a danger. Unfortunately MLSD was not designed with recursion in
mind. (Breadth-first traversal could be problematic due to lack of
sufficient navigation information.)
Of course all of Kermit's other MGET switches can be used too, e.g. for
finer-grained file selection (by date, size, etc), for moving or
renaming files as they arrive, to override Kermit's automatic per-file
text/binary mode switching, to pass the incoming files through a
filter, to convert text-file character sets, and so on.
3.11.4.4. NLST/MLSD Summary Table
Here's a table summarizing MGET behavior when the server supports both
NLST and MLSD. /NLST and /MLSD switches are included for clarity to
indicate which protocol is being used, and the expected effects. In
practice you can omit the /NLST and /MLSD switches and the Kermit
client chooses the appropriate or desired protocol as described above.
Sample commands presume a Unix file system on the server, but of course
the server can have any file system or syntax at all.
User's Command FTP Sends Remarks
mget /nlst NLST Gets a list of all the files in the server's current
and downloads each file. The list includes names only, so Kermit also
must send SIZE and MDTM directives if size and timestamp information is
required (this is always true of NLST). Sending NLST without an
argument is allowed by the RFC959 NLST definition and by the Kermit FTP
client, but might not work with other clients, and also might not work
with every server.
mget /nlst foo NLST foo If "foo" is a directory, this gets a list of
all the files from the server's "foo" directory and downloads each
file; otherwise this downloads the file named "foo" (if any) from the
server's current directory.
mget /nlst *.txt NLST *.txt Gets a list of the files in the server's
current directory whose names match the pattern *.txt, and then
downloads each file from the list. Because we are using NLST, we send
the filespec (*.txt) to the server and the server interprets any
wildcards.
mget /nlst foo/*.txt NLST foo/*.txt Gets a list of the files in the
server's "foo" directory whose names match the pattern *.txt, and then
downloads each file from the list (server interprets wildcards).
mget /nlst /match:*.txt NLST Gets a list of all the files in the
server's current directory and then downloads each one whose name
matches the pattern *.txt (client interprets wildcards).
mget /nlst /match:*.txt foo NLST foo Gets a list of all the files in
the server's "foo" directory and then downloads each one whose name
matches the pattern *.txt (client interprets wildcards).
mget /mlsd MLSD Gets a list of all the files from the server's current
directory and then downloads each one. The list might include size and
timestamp information, in which case Kermit does not need to send SIZE
and MDTM directives for each file (this is always true of MLSD).
mget /mlsd foo MLSD foo Gets a list of all the files from the server's
"foo" directory (where the string "foo" does not contain wildcards) and
then downloads each one. If "foo" is a regular file and not a
directory, this command is supposed to fail, but some servers have been
observed that send the file.
mget /mlsd *.txt MLSD Gets a list of all the files from the server's
current directory and then downloads only the ones whose names match
the pattern "*.txt". Because we are using MLSD and the MGET filespec is
wild, we do not send the filespec to the server, but treat it as though
it had been given in a /MATCH: switch and use it locally to match the
names in the list.
mget /mlsd foo/*.txt MLSD This one won't work because MLSD requires
that the notions of server directory and filename-matching pattern be
separated. However, the client, which can't be expected to know the
server's file-system syntax, winds up sending a request that the server
will (or should) reject.
mget /mlsd /match:*.txt MLSD Gets a list of all the files from the
server's current directory and then downloads only the ones whose names
match the pattern "*.txt" (client interprets wildcards).
mget /mlsd /match:*.txt foo MLSD foo If "foo" is a directory on the
server, this gets a list of all the files from the server's "foo"
directory and then downloads only the ones whose names match the
pattern "*.txt" (client interprets wildcards). This leaves the server
CD'd to the "foo" directory; there's no way the client can restore the
server's original directory because MLSD doesn't give that information,
and since the client can not be expected to know the server's
file-system syntax, it would not be safe to guess. If "foo" is a
regular file, MLSD fails.
mget /mlsd foo bar MLSD This one is problematic. You're supposed to be
able to give MGET a list a filespecs; in this case we name two
directories. The client must change the server's directory to "foo" to
get the list of files, and then the files themselves. But then it has
no way to return to the server's previous directory in order to do the
same for "bar", as explained in the previous example.
mget /mlsd /match:* [abc] MLSD [abc] Including a /MATCH: switch forces
[abc] to be sent to the server even though the client would normally
think it was a wildcard and hold it for local interpretation. In this
example, [abc] might be a VMS directory name.
mget /mlsd /match:* t*.h MLSD t*.h Contrary to the MLSD specification,
some MLSD-capable FTP servers do interpret wildcards. This form of the
MGET command can be used to force a wildcard to be sent to the server
for interpretation.
When MLSD is used implicitly (that is, without an /MLSD switch given to
force the use of MLSD) and an MGET command such as "mget foo/*.txt"
fails, Kermit automatically falls back to NLST and tries again.
3.11.5. References
1. Postel, J., and J. Reynolds, File Transfer Protocol (FTP), RFC 959,
October 1985: [372]ftp://ftp.isi.edu/in-notes/rfc959.txt.
2. Hethmon, P, and R. Elz, Feature negotiation mechanism for the File
Transfer Protocol, RFC 2389, August 1998:
[373]ftp://ftp.isi.edu/in-notes/rfc2389.txt.
3. Elz, R, and P. Hethmon, Extensions to FTP, Internet Draft
draft-ietf-ftpext-mlst-16.txt, September 2002:
[374]http://www.ietf.org/internet-drafts/draft-ietf-ftpext-mlst-16.
txt.
4. [375]The Kermit FTP Client (overview).
[ [376]Top ] [ [377]FTP Top ] [ [378]C-Kermit Home ] [ [379]Kermit Home
]
4. FILE SCANNING
A new feature called file scanning is used in various contexts to
determine if a file is text or binary, and if it is text, what kind of
text. The overhead of file scanning is surprisingly tolerable, usually
about a quarter second per file. File scanning is now used instead of
filename patterns unless you SET FILE SCAN OFF, which restores the
previous behavior.
The primary benefit of file scanning is in file transfer. For all
practical purposes, now you can stop worrying about whether a file
should be sent in binary or text mode, or about sending mixtures of
text and binary files in a single operation, or configuring and
fine-tuning your lists of binary-file and text-file name patterns: now
it all just works.
File scanning is done by the file sender, which determines the type of
each file before it sends it and informs the receiver (Kermit or FTP
server) of the type. File scanning is NOT done by the receiver, because
it is the sender's responsibility to determine each file's type, send
the file in the right mode, and inform the receiver of the mode. If
both transfer partners are capable of this (or any other) form of
automatic text/binary mode switching, then files can be sent in both
directions with no worries about corruption due to inappropriate
transfer mode. (As noted in [380]Section 3, FTP servers don't do this,
so this discussion does not apply when using Kermit to download from an
FTP server.)
The rest of this section is mainly for the curious. If you don't read
it and simply accept all defaults, every file you send should go in the
appropriate mode automatically. As always, however, for character-set
translation to work for 7- and 8-bit character-set files, the
appropriate SET FILE CHARACTER-SET command(s) must have been executed
to identify their encoding (Kermit's default file character-set is
neutral ASCII except on platforms like HP-UX or DG/UX, where the
default file character-set is known). And of course, receiving is
another matter -- obviously the other Kermit must also send each file
in the appropriate mode.
Scanning is more reliable than filename patterns simply because
filenames are not reliable indicators of the file's contents. Classic
examples include ".doc" files, which are binary if Microsoft Word
documents but text on most other platforms, and ".com" files, which are
binary on DOS and Windows but text on VMS. Anyway, nobody knows the
naming conventions (if any) of all the applications (and persons!) on
your computer. Scanning, on the other hand, determines each file's type
by inspecting its contents rather than just looking at its name.
Also, file patterns -- even when they work as intended -- categorize
each file only as text or binary, whereas file scanning can make finer
distinctions:
BINARY
Binary data, not to be converted in any way. Examples include
binary machine code (executable programs), graphics images (GIF,
JPG, etc), compressed files (Z, GZ, etc), archives and packages
(ZIP, TAR, RPM, etc), object files and libraries (OBJ, DLL,
etc).
7-BIT TEXT
Text encoded in a 7-bit character set such as ASCII or one of
the ISO 646 national versions. Kermit has no way to tell which
character is used, only that it's 7-bit text. Typical examples
include program source code, README files, Perl or Kermit
scripts, plain-text email, HTML, TeX, and various textual
encodings of binary files: Hex, Base64, etc. When sending such
files, the FILE DEFAULT 7BIT-CHARACTER-SET is used as the file
character-set, and then the appropriate transfer character set
is chosen from the associations list (ASSOCIATE, SHOW
ASSOCIATIONS).
8-BIT TEXT
Text encoded in an 8-bit character set such as Latin-1, Latin-2,
Latin/Hebrew, Latin/Cyrillic, KOI8, HP-Roman8, JIS X 0208, Code
Page 437, or Code Page 1252. Again, Kermit has no way of knowing
which particular set is in use, only that it's 8-bit text. When
sending such files, the FILE DEFAULT 8BIT-CHARACTER-SET is used
as the file character-set, and then the appropriate transfer
character set is chosen from the associations list.
UCS2 TEXT
Unicode in its basic form, 16 bits (2 octets) per character.
When sending such files, UCS2 is the file character-set and the
byte order is identified automatically; the appropriate transfer
character set is chosen from the associations list. Normally
this would be UTF8. UTF-16 is not supported yet; Kermit's
Unicode translations are restricted to Plane 0, the Base
Multilingual Plane (BMP).
UTF8 TEXT
Unicode in its 8-bit transformation format. When sending such
files, UTF8 is the file character-set; the appropriate transfer
character set is chosen from the associations list, normally
UCS2 or UTF8.
File scanning is available in UNIX C-Kermit, in K-95, and to a limited
extent, in VMS C-Kermit (full scanning is problematic in VMS because
even plain-text files might contain binary record-format information).
The relevant commands are:
SET TRANSFER MODE { AUTOMATIC, MANUAL }
Tells whether the file-transfer mode (text or binary) should be
set by automatic or "manual" means. AUTOMATIC is the default,
which allows any of the automatic methods that are enabled to do
their jobs: FILE SCAN, FILE PATTERNS, peer recognition, etc.
MANUAL lets you control the transfer mode with the SET FILE TYPE
commands. As always, /TEXT and /BINARY switches on your
file-transfer commands override all other methods; if you give
one of these switches, scanning is not done. SHOW TRANSFER
displays the current TRANSFER MODE setting.
SET FILE SCAN { ON [ number ], OFF }
Turns this feature on and off. It's ON by default. When OFF, the
previous rules apply (SET FILE PATTERNS, etc). When ON is given,
you can also specify a number of bytes to be scanned. The
default is 49152 (= 48K). If a negative number is given, the
entire file is scanned, no matter how big, for maximum certainty
(for example, a PostScript file that appears to be plain text
might include an embedded graphic past the normal scanning
limit). SHOW FILE displays the current FILE SCAN setting.
SET FILE DEFAULT 7BIT-CHARACTER-SET name
Tells the 7-bit character-set to use if scanning identifies a
7-bit text file, e.g. GERMAN. SHOW FILE displays the current SET
FILE DEFAULT settings. So does SHOW CHARACTER-SETS.
SET FILE DEFAULT 8BIT-CHARACTER-SET name
Tells the 8-bit character-set to use if scanning identifies an
8-bit text file, e.g. LATIN1. SHOW FILE and SHOW CHARACTER-SET
display this.
ASSOCIATE FILE-CHARACTER-SET fcs tcs
When sending files and a file character-set (fcs) is identified
by scanning, this tells C-Kermit which transfer character-set
(tcs) to translate it to. It also allows C-Kermit to set the
appropriate transfer character-set automatically whenever you
give a SET FILE CHARACTER-SET command.
ASSOCIATE TRANSFER-CHARACTER-SET tcs fcs
When receiving files and a file arrives whose transfer
character-set (tcs) is announced by the sender, this command
tells C-Kermit which file character-set (fcs) to translate it
to. It also allows C-Kermit to set the appropriate file
character-set whenever you give a SET TRANSFER CHARACTER-SET
command.
SET FILE CHARACTER-SET name
When given for a 7-bit set, also sets FILE DEFAULT
7BIT-CHARACTER-SET to the same set. When given for an 8-bit set,
also sets FILE DEFAULT 8BIT-CHARACTER-SET to the same set. If an
ASSOCIATE FILE-CHARACTER-SET command has been given for this
set, also sets the corresponding transfer character-set.
DIRECTORY /XFERMODE [ filespec ]
Performs a file scan of the given files, listing the result for
each file. If FILE SCAN is OFF but PATTERNS are ON, the result
shown according to the current FILE TEXT-PATTERNS and
BINARY-PATTERNS, and are restricted to (B) and (T). When FILE
SCAN is ON, the results are:
(B) Binary
(T)(7BIT) Text: 7-bit
(T)(8BIT) Text: 8-bit
(T)(UTF8) Text: Unicode UTF8
(T)(UCS2BE) Text: Unicode UCS2 Big Endian
(T)(UCS2LE) Text: Unicode UCS2 Little Endian
So you can use DIR /XFER to get a preview of how each file in a
selected group will be transferred. Everything to the right of
the (B) or (T) is new. If FILE SCAN is OFF, you only get the (B)
or (T) as before.
Note: Big and Little Endian refer to the ordering of bytes
within a computer word. Big Endian architecture is standard and
is used on most non-PC computers. Little Endian architecture is
used on PCs.
To illustrate file-transfer with scanning, suppose you have a directory
containing a mixture of text and binary files, and each text file can
be 7-bit German ISO 646, 8-bit Latin-1, or Unicode in any of the
following forms: UCS2 Little Endian, UCS2 Big Endian, or UTF8
([381]UTF-16 is not supported yet). Assuming all the built-in defaults
are in effect, the following three commands do the job:
set file char german ; This sets the default for 7-bit text files
set file char latin1 ; This sets the default for 8-bit text files
send *
Each file is sent in the appropriate mode (text or binary), with text
files converted to the appropriate transfer character-set and labeled
so the receiver can convert them according to its own local
conventions.
By the way, what if you want to inhibit character-set translation but
still allow automatic text/binary mode switching? Previously, you could
simply SET TRANSFER CHARACTER-SET TRANSPARENT. But now with file
scanning, the file and transfer character-sets are set automatically
per file. A new command was added for this purpose:
SET TRANSFER TRANSLATION { ON, OFF }
Enables and disables file-transfer character-set translation. It
is enabled by default.
When TRANSFER TRANSLATION is OFF but FILE SCAN is ON, files are still
scanned to see if they are text or binary, but no character-set
translation is done when they text: only the normal record-format
conversion.
Like all SET commands, SET TRANSFER TRANSLATION is global and
persistent. You can also force a particular file-transfer command
(SEND, MSEND, GET, RECEIVE, TRANSMIT, etc) to not translate without
affecting the global translation settings by including the new
/TRANSPARENT switch, e.g.
send /transparent oofa.txt
As of C-Kermit 8.0.206, SET TRANSFER CHARACTER-SET TRANSPARENT implies
SET TRANSFER TRANSLATION OFF.
File scanning is also used in the TYPE command. The source file type
and character set are determined as above, and then the file is
automatically converted to your display character-set, line by line. In
Kermit 95, the display character-set is Unicode, perhaps converted to
your current console code page; in other versions of C-Kermit, it is
your current file character-set. Thus if you have the following set
appropriately:
SET FILE CHARACTER-SET (necessary in Unix but not K95)
SET FILE DEFAULT 7BIT CHARACTER-SET
SET FILE DEFAULT 8BIT CHARACTER-SET
then you should be able to TYPE any text file and see something
reasonable. For example, in Unix, if your DEFAULT 7BIT-CHARACTER-SET is
ITALIAN and your DEFAULT 8BIT-CHARACTER-SET is LATIN1, and your FILE
CHARACTER-SET is LATIN1, you can TYPE an Italian ISO 646 file, a
Latin-1 file, or any kind of Unicode file, and have it translated
automatically to Latin-1 for your display.
In the GUI version of Kermit 95, you can see mixtures of many different
scripts if the file is UTF8 or UCS2: Roman, Cyrillic, Hebrew, Greek,
Armenian, Georgian, etc, all on the same screen at once.
File scanning also adds a new criterion for file selection, i.e. to
select only text (or binary) files. Several commands now include a new
switch, /TYPE:{BINARY,TEXT,ALL}. BINARY means select only binary
regular files (not directories). TEXT means select only text files. ALL
means don't scan; select all files. Examples:
SEND /TYPE:BINARY *.*
Sends only binary files, skipping over text files.
NOTE: File scanning is NOT done when using external protocols (because
the external protocol programs, such as sz, are processing each file,
not Kermit).
DIRECTORY /TYPE:TEXT
Lists only text files but not binary files.
DELETE /TYPE:BINARY foo.*
Deletes all foo.* files that are regular binary files but does
not delete any text files.
CHMOD /TYPE:BINARY 775 *
(UNIX) Changes the permissions of all binary files to 775.
When FILE SCAN is OFF and FILE PATTERNS are ON, behavior is as before
with PATTERNS ON, but with some improvements:
* Pathnames are now stripped prior to pattern matching.
* Backup suffixes (like .~3~) are stripped prior to pattern matching.
[ [382]Top ] [ [383]Contents ] [ [384]C-Kermit Home ] [ [385]Kermit
Home ]
5. FILE AND DIRECTORY NAMES CONTAINING SPACES
Prior to the introduction of the graphical user interface (GUI), it was
inconceivable that file or directory names could contain spaces,
because space is a field delimiter in all command languages. GUIs,
however, use dialog boxes for filenames, so there is never any question
of distinguishing a filename from adjacent fields -- because there are
no adjacent fields -- and therefore it has become quite common on
computers that have GUIs to have file and directory names composed of
multiple words. Of course this poses problems for command shells and
other text-oriented programs.
Most command shells address these problems by allowing such names to be
enclosed in doublequotes, e.g.:
cd "c:\Program Files"
C-Kermit previously used braces for this:
cd {c:\Program Files}
which was not what most people expected. And even when braces were
used, Kermit had difficulties with completion, file menus, and so
forth, within braced fields.
C-Kermit 8.0 allows either doublequotes or braces to be used for
grouping:
send "this file"
send {this file}
rename "this file" "that file"
rename {this file} "that file"
rename "this file" {that file}
cd {Program Files}
cd "Program Files"
Note that the doublequotes or brackets must enclose the whole file or
directory specification:
"c:\My Directory"
not:
c:\"My Directory"
In C-Kermit 8.0, you can also use completion on these filenames, in
which case Kermit supplies the quotes (or braces) automatically.
Example (in which the current directory contains only one file whose
name starts with "th" and its full name is "this file" (without the
quotes, but with the space)):
cat th<Tab>
Kermit repaints the filename field like this:
cat "this file"
That is, it backspaces over the original "th" and then writes the
filename in doublequotes.
If completion is only partial, Kermit still supplies the quotes, but in
this case also beeps. To continue the filename, you must first
backspace over the closing quote. The closing quote is supplied in this
case to make sure that you can see the spaces, especially if they are
trailing. For example, if the current directory contains two files
whose names start with "th", and their fill names are "this file" and
"this other file":
cat th<Tab>
Kermit prints:
cat "this "<Beep>
If it didn't print the closing quote, you would probably wonder why it
was beeping.
Also, if you begin a filename field with a doublequote or opening
brace, now you can use completion or get ?-help; this was never
possible before.
C-Kermit>type "thi? Input file specification, one of the following:
this file this other file
C-Kermit>type "thi_
[ [386]Top ] [ [387]Contents ] [ [388]C-Kermit Home ] [ [389]Kermit
Home ]
6. OTHER COMMAND PARSING IMPROVEMENTS
6.1. Grouping Macro Arguments
Doublequotes now can be used in macro invocations to group arguments
containing spaces, where previously only braces could be used:
define xx show args
xx one "this is two" three
Result:
Macro arguments at level 0 (\v(argc) = 4):
\%0 = xx
\%1 = one
\%2 = this is two
\%3 = three
Also, you can now quote braces and quotes in macro args (this didn't
work before). Examples:
xx "{" ; The argument is a single left brace
xx {"} ; The argument is a doublequote character
In case this new behavior interferes with your scripts, you can restore
the previous behavior with:
SET COMMAND DOUBLEQUOTING OFF
6.2. Directory and File Name Completion
C-Kermit 8.0 also includes better completion for directory names, e.g.
in the CD command. If the name typed so far uniquely matches a
directory name, it is completed (as before), but now if the directory
contains any subdirectories, completion is partial (allowing you to
supply additional path segments without backspacing); otherwise it is
complete.
Completion has also been improved for file and directory names that
contain not only spaces (as described above) but also "metacharacters"
such as asterisk (*) and tilde (~): now the field is repainted if
necessary. For example, if the current directory contains only one file
whose name contains "blah", then in:
type *blah<Tab>
"*blah" is replaced by the filename. In earlier releases, the part
typed so far was left on the command line (and in the history buffer),
so even when the original command worked, the recalled version would
not. Similarly for ~ (the nearly-universal Unix notation for username):
type ~olga/x<Tab>
is repainted as (e.g.):
type /users/home/olga/x(Beep)
Speaking of command history, the new SHOW HISTORY command shows your
command history and recall buffer. SAVE COMMAND HISTORY saves it into a
file of your choice.
6.3. Passing Arguments to Command Files
The method for passing arguments to command files has been improved.
Prior to C-Kermit 7.0 there was no provision for doing this. In
C-Kermit 7.0, the TAKE command was changed to allow arguments to be
given after the filename:
take commandfile arg1 arg2 ...
This was accomplished by replacing the current \%1, \%2, etc, with the
given arguments, since a new set of macro argument variables is created
only when a macro is executed, not a command file. It is much more
intuitive, however, if arguments to command files worked like those to
macros: the command file sees the arguments as its own \%1, \%2, etc,
but the caller's variables are not disturbed. C-Kermit 8.0 accomplishes
this by automatically creating an intermediate temporary macro to start
the command file (if any arguments were given), thus creating a new
level of arguments as expected.
6.4. More-Prompting
The familiar --more?-- prompt that appears at the end of each screenful
of command-response output now accepts a new answer: G (Go) meaning
"show all the rest without pausing and asking me any more questions". P
(Proceed) is a synonym for G.
6.5. Commas in Macro Definitions
As noted in the [390]C-Kermit manual, comma is used to separate
commands in a macro definition. Even when the macro is defined on
multiple lines using curly-brace block-structure notation without
commas, the definition is still stored internally as a comma-separated
list of commands. Therefore special tricks are needed to include a
comma in a command. The classic example is:
define foo {
(some command)
if fail echo Sorry, blah failed...
}
This would result in Kermit trying to execute a "blah" command. This
could always be handled by enclosing the text in braces:
define foo {
(some command)
if fail echo {Sorry, blah failed...}
}
but doublequotes (more intuitive) should have worked too. Now they do:
define foo {
(some command)
if fail echo "Sorry, blah failed..."
}
6.6. Arrow Keys
As of version 8.0.201, C-Kermit on most platforms lets you access the
command history buffer with arrow keys, just as you always could with
control characters. The restrictions are:
1. Only Up and Down arrow keys are accepted.
2. Only 7-bit ANSI arrow-key sequences are understood (ESC followed by
[ or uppercase letter O, followed by uppercase letter A or (up) B
(down).
This change was made to facilitate command recall in Linux-based PDAs
that don't have a Control key, or at least not one that's easily (or
always) accessible, such as the Sharp Zaurus SL5500.
[ [391]Top ] [ [392]Contents ] [ [393]C-Kermit Home ] [ [394]Kermit
Home ]
7. NEW COMMANDS AND SWITCHES
See [395]Section 4 for more about file scanning and the /TYPE: switch.
ASK[Q] [ /TIMEOUT:number /QUIET /DEFAULT:text ] variable [ prompt ]
The new optional /TIMEOUT: switch for ASK and ASKQ causes the
command to time out and and fail if no response is given within
the specified number of seconds, 1 or greater (0 or less means
no timeout, wait forever). This works just like SET ASK-TIMER,
except its effect is local to the ASK command with which it is
given and it does not disturb the global ask timer setting. The
new /QUIET switch tells Kermit not to print an error message if
the ASK or ASKQ command times out waiting for a response.
Version 8.0.211 adds the /DEFAULT:text switch for ASK-Class
commands (ASK, ASKQ, and GETOK). This lets you supply a default
answer in case the user supplies an empty answer or the
/TIMEOUT: switch was included and the time limit expired without
an answer. In both these cases, the command succeeds.
CAT filename
Equivalent to TYPE /NOPAGE.
CDUP
Changes Kermit's local working directory to the parent of the
current one. Equivalent to "cd .." in UNIX or Windows, "cd [-]"
in VMS, "cd ^" in AOS/VS, etc; in other words, it's a
platform-independent way of moving one level up in a directory
tree.
CHMOD [ switches ] permission files
UNIX only. Sets file permissions for one or more files or
directories. The permission must be given as an octal number,
e.g. 664, 755. Switches: /DIRECTORIES, /FILES, /NOLIST, /PAGE,
/DOTFILES, /LIST, /NOPAGE, /RECURSIVE, /TYPE:{TEXT,BINARY,ALL},
/SIMULATE. The /TYPE: switch allows selection of only text or
binary files. For example, if you have a mixture of source files
and executables, you can use "chmod /files /type:text 664" to
give owner/group read/write and world read permission to the
text files, and "chmod /files /type:binary 775" to give the same
plus execute permission to the executables. Use /SIMULATE to see
which files would be affected, without actually changing their
permissions.
CLEAR KEYBOARD-BUFFER
Flushes any as-yet unread characters from the keyboard input
buffer. Useful for flushing typeahead in scripts.
CONTINUE
When given at an interactive command prompt that was reached by
issuing a PROMPT command (described in this section) from a
script, this command returns to the script, continuing its
execution at the command after the PROMPT command. In this
context, CONTINUE is simply a more-intuitive synonym for END.
COPY, RENAME, and TRANSLATE
These commands now work on file groups if the target filename is
a directory, e.g. "copy oofa.* ..", "rename * ~olga/tmp/"
COPY /APPEND source destination
The source file specification can now include wildcards, in
which case all of the source files that match will go into the
destination file in alphabetical order by name.
DELETE /ASK
Asks permission to delete each file before deleting it. In
C-Kermit 7.0, the answers were "yes" (or "ok") and "no".
C-Kermit 8.0 adds "go" (meaning, delete all the rest without
asking) and "quit" (cancel the DELETE command and return to the
prompt).
DELETE /DIRECTORIES
Deletes not only files but also directories.
DELETE /RECURSIVE
Deletes all files that match the given file specification in the
current (or given) directory and all directories beneath it.
DELETE /SUMMARY
Prints only the number of files deleted and total size freed,
without listing each file.
DELETE /TREE
Shorthand for DELETE /RECURSIVE /DIRECTORIES /DOTFILES/.
Equivalent to Windows DELTREE or Unix "rm -Rf". If no file
specification is given, the contents of the current directory,
plus all of its subdirectories and their contents, are deleted.
DELETE /TYPE:BINARY
Delete only regular binary files (requires FILE SCAN ON).
DELETE /TYPE:TEXT
Delete only regular text files (requires FILE SCAN ON).
DIRECTORY [ switches ] [ filespec [ filespec [ filespec ... ] ] ]
The DIRECTORY command now accepts more than one file
specification; e.g. "directory moon.txt sun.doc stars.*".
DIRECTORY /NORECURSIVE xxx
If xxx is a directory name, forces listing of the directory
itself rather than its contents.
DIRECTORY /FOLLOWLINKS xxx
(UNIX only) Tells the DIRECTORY command to follow symbolic
links. This not the default because it can cause endless loops.
DIRECTORY /NOFOLLOWLINKS xxx
(UNIX only) Tells the DIRECTORY command not to follow symbolic
links, but rather, merely to list them. This is the default.
DIRECTORY /OUTPUT:filename
Sends the results of the DIRECTORY command to the given file.
DIRECTORY /SUMMARY
Prints only the number of directories and files and the total
size, without listing each file.
DIRECTORY /TYPE:{TEXT,BINARY}
Shows only files of the selected type, based on file scan.
DIRECTORY /XFERMODE
Now shows results of file scan (see [396]Section 4).
FOPEN [ switches ] channel filename
As of version 8.0.211, FOPEN allows /dev/tty as a filename in
Unix-based operating systems.
FREAD /TRIM
(8.0.211) Trims any trailing blanks or tabs from the item (such
as a line of text) that it has read.
FREAD /UNTABIFY
(8.0.211) Converts Horizontal Tab characters to the appropriate
number of spaces, based on VT100-like tab stops (1,9,17,25,...).
GREP [ switches ] pattern files
Similar to Unix grep command: displays file lines that match the
given [397]pattern. Switches:
/COUNT[:variable]
Don't show the matching lines, just tell how many lines
match. If a variable name is specified, the count is
stored in the given variable.
/DOTFILES
Include files whose names begin with dot.
/LINENUMBERS
Show line numbers of matching lines.
/NAMEONLY
only list the names of files that contain matching lines,
but not the lines themselves.
/NOBACKUP
Skip backup files.
/NOCASE
Ignore alphabetic case while pattern matching.
/NODOTFILES
skip files whose names start with dot (period).
/NOLIST
Suppress output but set SUCCESS or FAILURE according to
search result.
/NOMATCH
Look for lines that do not match the pattern.
/NOPAGE
Don't pause between screens of output.
/OUTPUT:filename
Write results into the given file.
/PAGE
Pause between screens of output.
/RECURSIVE
Search files in subdirectories too.
/TYPE:{TEXT,BINARY}
Search only files of the specified type.
Synonyms: FIND, SEARCH.
GETOK /TIMEOUT:n /QUIET /DEFAULT:text
The new /QUIET switch instructs GETOK, when given a timeout, not
to print an error message if it times out. As of 8.0.211, a
default answer can be supplied (see ASK).
HEAD [ switches ] filename
Equivalent to TYPE /HEAD [ other-switches ] filename.
HELP DATE
Explains date-time formats, including timezone notation and
delta times.
HELP FIREWALLS
Explains the firewall negotiation capabilities of your version
of Kermit.
KCD [ symbolic-directory-name ]
Changes Kermit's working directory to the named symbolic
directory, such as such as exedir, inidir, startup, download, or
and home. Type "kcd ?" for a list of symbolic directory names
known to your copy of Kermit, or give the new ORIENTATION
command for a more detailed explanation. If you give a KCD
command without a directory name, Kermit returns to its "home"
directory, which is determined in some way that depends on the
underlying operating system, but which you can redefine with the
(new) SET CD HOME command. Your home directory is shown by SHOW
CD and it's also the value of the \v(home) variable.
LICENSE
Displays the C-Kermit license.
L-commands
When Kermit has a connection to a Kermit or FTP server, file
management commands such as CD, DIRECTORY, and DELETE might be
intended for the local computer or the remote server. C-Kermit
8.0.200 and earlier always executes these commands on the local
computer. If you want them executed by the remote server, you
have to prefix them with REMOTE (e.g. REMOTE CD) or use special
R-command aliases (e.g. RCD = REMOTE CD, RDIR = REMOTE DIR,
etc). But this feels unnatural to FTP users, who expect
unprefixed file management commands to be executed by the remote
server, rather than locally. C-Kermit 8.0.201 adds automatic
locus switching to present an FTP-like interface for FTP
connections and the normal Kermit interface for Kermit
connections, and a SET LOCUS command (described below) to
control whether or how this is done. For when LOCUS is REMOTE, a
new set of commands was added for local management: LCD (Local
CD), LDIR (Local DIR), etc. These are described below under SET
LOCUS.
MORE filename
Equivalent to TYPE /PAGE.
ORIENTATION
Displays symbolic directory names and the corresponding variable
names and values. The symbolic names, such as exedir, inidir,
startup, download, and home, can be used as arguments to the new
KCD command.
PROMPT [ text ]
For use in a macro or command file: enters interactive command
mode within the current context ([398]Section 8.1). If the
optional text is included, the prompt is set to it. The text can
include variables, functions, etc, as in the SET PROMPT command.
They are evaluated each time the prompt is printed. Unlike the
SET PROMPT command, the text argument applies only to the
current command level. Thus you can have different prompts at
different levels.
REMOTE SET MATCH { DOTFILE, FIFO } { ON, OFF }
Allows the client to tell the server whether wildcards sent to
the server should match dot files (files whose names begin with
period) or FIFOs (named pipes). See SET MATCH.
SET ATTRIBUTE RECORD-FORMAT { ON, OFF }
Allows control of the Kermit's Record-Format attribute. Set this
to OFF in case incoming file are refused due to unknown or
invalid record formats if you want to accept the file anyway
(and, perhaps, postprocess it to fix its record format).
SET CD HOME [ directory ]
Specifies the target directory for the CD and KCD commands, when
they are given without an argument, and also sets the value of
the \v(home) variable.
SET EXIT HANGUP { OFF, ON }
Normally ON, meaning that when Kermit exits, it also explicitly
hangs up the current SET LINE / SET PORT serial port according
to the current SET MODEM TYPE and SET MODEM HANGUP METHOD, and
closes the port device if it was opened by Kermit in the first
place (as opposed to inherited). SET EXIT HANGUP OFF tells
Kermit not to do this. This can't prevent the operating system
from closing the device when Kermit exits (and it's a "last
close") but if the port or modem have been conditioned to
somehow ignore the close and keep the connection open, at least
Kermit itself won't do anything explicit to hang it up or close
it.
SET FILE EOF { CTRL-Z, LENGTH }
Specifies the end-of-file detection method to be used by
C-Kermit when sending and receiving text files, and in the TYPE
and similar text-file oriented commands. The normal and default
method is LENGTH. You can specify CTRL-Z when handling CP/M or
MS-DOS format text files, in which a Ctrl-Z (ASCII 26) character
within the file marks the end of the file.
SET FILE LISTSIZE number
Allocates space for the given number of filenames to be filled
in by the wildcard expander. The current number is shown by SHOW
FILE. If you give a command that includes a filename containing
a wildcard (such as "*") that matches more files that Kermit's
list has room for, you can adjust the list size with this
command.
SET FILE STRINGSPACE number
Allocates space for the given amount of filename strings for use
by the wildcard expander. The current number is shown by SHOW
FILE. The number is the total number of bytes of all the file
specifications that match the given wildcard.
If you need to process a bigger list of files than your computer has
memory for, you might be able use an external file list. The Kermit
SEND and the FTP PUT and GET commands accept a /LISTFILE: switch,
which gives the name of a file that contains the list of files to be
transferred. Example for UNIX:
!find . -print | grep / > /tmp/names
ftp put /update /recursive /listfile:/tmp/names
SET LOCUS { AUTO, LOCAL, REMOTE }
Added in C-Kermit 8.0.201. Sets the locus for unprefixed file
management commands such as CD, DIRECTORY, MKDIR, etc. When
LOCUS is LOCAL these commands act locally and a REMOTE (or R)
prefix (e.g. REMOTE CD, RCD, RDIR) is required to send file
management commands to a remote server. When LOCUS is REMOTE, an
L prefix is required to issue local file management commands
(e.g. LCD, LDIR). The word LOCAL can't be used as a prefix since
it is already used for declaring local variables. LOCUS applies
to all types of connections, and thus is orthogonal to SET
GET-PUT-REMOTE, which selects between Kermit and FTP for remote
file-transfer and management commands. The default LOCUS is
AUTO, which means we switch to REMOTE whenever an FTP connection
is made, and to LOCAL whenever a non-FTP connection is made, and
switch back accordingly whenever a connection is closed. So by
default, Kermit behaves in its traditional manner unless you
make an FTP connection, in which case it acts like a regular FTP
client (but better :-) LOCUS applies to the following
commands:
Unprefixed Remote Local Description
CD (CWD) RCD LCD Change (Working) Directory
CDUP RCDUP LCDUP CD Up
PWD RPWD LPWD Print Working Directory
DIRECTORY RDIR LDIR Request a directory listing
DELETE RDEL LDEL Delete (a) file(s)
RENAME RREN LREN Rename a file
MKDIR RMKDIR LMKDIR Create a directory
RMDIR RRMDIR LRMDIR Remove a directory
SET MATCH { DOTFILE, FIFO } { ON, OFF }
Whether C-Kermit filename patterns (wildcards) should match
filenames that start with dot (period), or (Unix only) FIFOs
(named pipes). The defaults are to skip dotfiles in Unix but
match them elsewhere, and to skip FIFOs. Applies to both
interactive use and to server mode, when the server receives
wildcards, e.g. in a GET command. Also see REMOTE SET MATCH.
SET OPTIONS DIRECTORY /DOTFILES
Now works for server listings too (UNIX only). Give this command
prior to having Kermit enter server mode, and then it will show
files whose names begin with dot (period) when sent a REMOTE
DIRECTORY command.
SET QUIET ON
(as well as the -q command-line option) Now applies also to:
+ SET HOST connection progress messages.
+ "Press the X or E key to cancel" file-transfer message.
+ REMOTE CD response.
+ REMOTE LOGIN response.
SET RECEIVE PERMISSIONS { ON, OFF }
Tells C-Kermit whether to set the permissions of incoming files
(received with Kermit protocol) from the permissions supplied in
the file's Attribute packet (if any). Normally ON. Also see SET
SEND PERMISSIONS.
SET ROOT directory
Like UNIX chroot, without requiring privilege. Sets the root for
file access, does not allow reference to or creation of files
outside the root, and can't be undone.
SET SEND PERMISSIONS { ON, OFF }
Tells C-Kermit whether to include file permissions in the
attributes it includes with each file when sending with Kermit
protocol. Also see SET RECEIVE PERMISSIONS.
SET TCP { HTTP-PROXY, SOCKS-SERVER } /USER:name /PASSWORD:text
These commands now allow specification of username and password.
SET TERMINAL . . .
(See [399]Section 12.)
SET TRANSFER MESSAGE [ text ]
Sets an initial text message to be displayed in the
file-transfer display. The transfer message is automatically
deleted once used, so must be set each time a message a desired.
Any variables in the message are evaluated at the time the SET
command is given. If the optional text is omitted, any transfer
message that is currently set is removed. Synonym: SET XFER MSG.
SHOW TRANSFER displays it if it has been set but not yet used.
SHOW COMMUNICATIONS
In C-Kermit 8.0, SHOW COMMUNICATIONS, when given in remote mode
(i.e. before any connection has been established), tells the
typical dialout device name for the particular platform on which
it's running (e.g. TXA0: for VMS, or /dev/cua0p0 for HP-UX). On
Unix platforms, it also tells the name of the lockfile
directory. This way, you have an idea of what the SET LINE
device name should look like, and if the SET LINE command fails,
you know the name of the directory or device that is protected
against you.
SHOW VARIABLES [ name [ name [ ... ] ] ]
In C-Kermit 8.0.201 you can request values of a list of built-in
(\v(xxx)) variables. Each name is a pattern, as before, but now
it a free pattern rather than an anchored one (explained in
[400]Section 8.12) so now "show var date time" shows the values
of all variables whose names include the strings "date" or
"time".
TAIL [ switches ] filename
Equivalent to TYPE /TAIL [ other-switches ] filename.
TRANSMIT /NOECHO [ other switches ] filename
The /NOECHO switch is equivalent to giving the command SET
TRANSMIT ECHO OFF prior to the TRANSMIT command, except the
switch affects only the command with which it was given and does
not affect the prevailing global setting.
TRANSMIT /NOWAIT [ other switches ] filename
The /NOWAIT switch is equivalent to giving the command SET
TRANSMIT PROMPT 0 prior to the TRANSMIT command, except the
switch affects only the command with which it was given and does
not affect the prevailing global setting.
TRANSMIT /NOWAIT /NOECHO /BINARY [ other switches ] filename
When the TRANSMIT command is given with the /NOWAIT, /NOECHO,
and /BINARY switches, this activates a special "blast the whole
file out the communications connection all at once" mode that
Kermit didn't have prior to version 8.0. There has been
increasing demand for this type of transmission with the advent
of devices that expect image (e.g. .JPG) or sound (e.g. .MP3)
files as raw input. The obvious question is: how does the
receiving device know when it has the whole file? This depends
on the device, of course; usually after a certain amount of time
elapses with nothing arriving, or else when Kermit hangs up or
closes the connection.
TYPE /CHARACTER-SET:name
Allows you to specify the character set in which the file to be
typed is encoded.
TYPE /NUMBER
Adds line numbers.
TYPE /OUTPUT:filename
Sends the results of the TYPE command to the given file.
TYPE /TRANSLATE-TO:name
Used in conjunction with TYPE /CHARACTER-SET:xxx; allows you to
specify the character set in which the file is to be displayed.
TYPE /TRANSPARENT
Used to disable character-set translation in the TYPE command,
which otherwise can take place automatically based on file
scanning, even when /CHARACTER-SET and /TRANSLATE-TO switches
are not given.
VOID text
Parses the text, evaluating any backslash items in it (such as
function calls) but doesn't do anything further, except possibly
printing error messages. Useful for invoking functions that have
side effects without using or printing their direct results,
e.g. "void \fsplit(\%a,&a)".
Symbolic Links in UNIX
The UNIX versions of C-Kermit have had /FOLLOWLINKS and /NOFOLLOWLINKS
switches added to several commands to control the treatment of symbolic
links. Different commands deal differently with symbolic links:
Kermit SEND, FTP MPUT
/NOFOLLOWLINKS is the default, which means symbolic links are
skipped entirely. The alternative, /FOLLOWLINKS, should be used
with caution, since an innocent link might point to a whole file
system, or it might cause a loop. There is no way in Kermit or
FTP protocol to send the link itself. We either skip them or
follow them; we can't duplicate them.
DIRECTORY
/NOFOLLOWLINKS is the default, which means the DIRECTORY command
lists symbolic links in a way that shows they are links, but it
does not follow them. The alternative, /FOLLOWLINKS, follows
links and gives information about the linked-to directories and
files.
DELETE, RMDIR
The DELETE command does not have link-specific switches. DELETE
never follows links. If you tell Kermit to delete a symbolic
link, it deletes the link itself, not the linked-to file. Ditto
for RMDIR.
COPY
The COPY command behaves just like the UNIX cp command; it
always follows links.
RENAME
The RENAME command behaves just like the UNIX mv command; it
operates on links directly rather than following.
[ [401]Top ] [ [402]Contents ] [ [403]C-Kermit Home ] [ [404]Kermit
Home ]
8. OTHER SCRIPTING IMPROVEMENTS
8.1. Performance and Debugging
A command cache for frequently used commands plus some related
optimizations increases the speed of compute-bound scripts by anywhere
from 50% to 1000%.
The new PROMPT command can be used to set breakpoints for debugging
scripts. If executed in a command file or macro, it gives you an
interactive command prompt in the current context of the script, with
all its variables, arguments, command stack, etc, available for
examination or change, and the ability to resume the script at any
point (END resumes it, Ctrl-C or STOP cancels it and returns to top
level).
The new Ctrl-C trapping feature ([405]Section 8.14) lets you intercept
interruption of scripts. This can be used in combination with the
PROMPT command to debug scripts. Example:
define ON_CTRLC {
echo INTERRUPTED BY CTRL-C...
echo The command stack has not yet been rolled back:
show stack
echo Type Ctrl-C again or use the END command to return to top level.
prompt Debug>
}
Adding this ON_CTRL definition to your script lets you interrupt it at
any point and get prompt that is issued at the current command level,
so you can query local variables, etc.
[ [406]Top ] [ [407]Contents ] [ [408]C-Kermit Home ] [ [409]Kermit
Home ]
8.2. Using Macros as Numeric Variables
A macro is a way to assign a value to a name, and then use the name to
refer to the value. Macros are used in two ways in Kermit: as
"subroutines" or functions composed of Kermit commands, which are
executed, or as variables to hold arbitrary values -- text, numbers,
filenames, etc.
When a macro is to be executed, its name is given as if it were a
C-Kermit command, optionally preceded by the word "do". When a macro is
used as a variable, it must be "escaped" with \m(xxx) (or equivalent
function, e.g. \s(xxx), \:(xxx), \fdefinition(xxx)), where xxx is the
macro name, for example:
define filename /usr/olga/oofa.txt
send \m(filename)
Of course variables can also hold numbers:
define size 17
declare \&a[\m(size)]
...
define index 3
if ( == \m(index) 3 ) echo The third value is: \&a[\m(index)]
evaluate index (\m(index) * 4)
if ( > \m(index) \m(size) ) echo Out of range!
But these are contexts in which only numbers are valid. C-Kermit 8.0
has been changed to treat non-escaped non-numeric items in strictly
numeric contexts as macro names. So it is now possible (but not
required) to omit the \m(...) notation and just use the macro name in
these contexts:
define size 17
declare \&a[size]
...
define index 3
if ( == index 3 ) echo The third value is: \&a[index]
evaluate index (index * 4)
if ( > index size ) echo Out of range!
This is especially nice for loops that deal with arrays. Here, for
example, is a loop that reverses the order of the elements in an array.
Whereas formerly it was necessary to write:
.\%n ::= \fdim(&a)
for \%i 1 \%n/2 1 {
.tmp := \&a[\%n-\%i+1]
.\&a[\%n-\%i+1] := \&a[\%i]
.\&a[\%i] := \m(tmp)
}
Recoding this to use macro names "i" and "n" instead of the backslash
variables \%i and \%n, we have:
.n ::= \fdim(&a)
for i 1 n/2 1 {
.tmp := \&a[n-i+1]
.\&a[n-i+1] := \&a[i]
.\&a[i] := \m(tmp)
}
which reduces the backslash count to less than half. The final
statement in the loop could be written ".\&a[i] ::= tmp" if the array
contained only numbers (since ::= indicates arithmetic expression
evaluation).
Also, now you can use floating-point numbers in integer contexts (such
as array subscripts), in which case they are truncated to an integer
value (i.e. the fractional part is discarded).
Examples of numeric contexts include:
* Array subscripts.
* Any numeric function argument.
* Right-hand side of ::= assignments.
* EVALUATE command or \fevaluate() function expression.
* The INCREMENT or DECREMENT by-value.
* IF =, >, <, !=, >=, and <= comparands.
* The IF number construct.
* FOR-loop variables.
* STOP, END, and EXIT status codes.
* The INPUT timeout value.
* PAUSE, WAIT, SLEEP, MSLEEP intervals.
* The SHIFT argument.
* Numeric switch arguments, e.g. TYPE /WIDTH:number, SEND
/LARGER:number.
* SCREEN MOVE-TO row and column number.
* Various SET DIAL parameters (timeout, retry limit, etc).
* Various SET SEND or RECEIVE parameters (packet length, window size,
etc).
* Various other SET parameters.
and:
* S-Expressions (explained in [410]Section 9).
Macro names used in numeric contexts must not include mathematical
operators. Although it is legal to create a macro called "foo+bar", in
a numeric context this would be taken as the sum of the values of "foo"
and "bar". Any such conflict can be avoided, of course, by enclosing
the macro name in \m(...).
[ [411]Top ] [ [412]Contents ] [ [413]C-Kermit Home ] [ [414]Kermit
Home ]
8.3. New IF Conditions
Several new IF conditions are available:
IF DECLARED arrayname
Explained in [415]Section 8.6.
IF KBHIT
Allows a script to test whether a key was pressed without
actually trying to read it.
IF KERBANG (Unix only)
True if Kermit was started from a Kerbang script. This is useful
for knowing how to interpret the \&@[] and \&_[] argument vector
arrays, and under what conditions to exit.
IF INTEGER n
This is just a synonym for IF NUMERIC, which is true if n
contains only digits (or, if n is a variable, its value contains
only digits).
By contrast, IF FLOAT n succeeds if n is a floating-point number OR an
integer (or a variable with floating-point or integer value).
Therefore, IF FLOAT should be used whenever any kind of number is
acceptable, whereas IF INTEGER (or IF NUMERIC) when only an integer can
be used.
[ [416]Top ] [ [417]Contents ] [ [418]C-Kermit Home ] [ [419]Kermit
Home ]
8.4. The ON_UNKNOWN_COMMAND Macro
The new ON_UNKNOWN_COMMAND macro, if defined, is executed whenever you
give a command that is not known to C-Kermit; any operands are passed
as arguments. Here are some sample definitions:
DEF ON_UNKNOWN_COMMAND telnet \%1 ; Treat unknown commands as hostnames
DEF ON_UNKNOWN_COMMAND dial \%1 ; Treat unknown commands phone numbers
DEF ON_UNKNOWN_COMMAND take \%1 ; Treat unknown commands as filenames
DEF ON_UNKNOWN_COMMAND !\%* ; Treat unknown commands as shell commands
The ON_CD macro, if defined, is executed whenever Kermit is given a CD
(change directory) command (8.0.211). Upon entry to this macro, the
directory has already changed and the new directory string is available
in the \v(directory) variable, and also as the first argument (\%1).
[ [420]Top ] [ [421]Contents ] [ [422]C-Kermit Home ] [ [423]Kermit
Home ]
8.5. The SHOW MACRO Command
The SHOW MACRO command has been changed to accept more than one macro
name:
(setq a 1 b 2 c 3)
show mac a b c
a = 1
b = 2
c = 3
An exact match is required for each name (except that case doesn't
matter). If you include wildcard characters, however, a pattern match
is performed:
show mac [a-c]*x
shows all macros whose names start with a, b, or c, and end with x.
[ [424]Top ] [ [425]Contents ] [ [426]C-Kermit Home ] [ [427]Kermit
Home ]
8.6. Arrays
A clarification regarding references to array names (as opposed to
array elements): You can use array-name "abbreviations" like &a only in
contexts that expect array names, like ARRAY commands or array-name
function arguments such as the second argument of \fsplit(). In a LOCAL
statement, however, you have to write \&a[], since "local &a" might
refer to a macro named "&a".
In function arguments, however, you MUST use the abbreviated form:
\fsplit(\%a,&a) or \fsplit(\%a,&a[]). If you include the backslash (as
in "\fsplit(\%a,\&a[])") a parse error occurs.
Here are the new array-related commands:
IF DECLARED arrayname
Allows a script to test whether an array has been declared. The
arrayname can be a non-array backslash variable such as \%1 or
\m(name), in which case it is evaluated first, and the result is
treated as the array name. Otherwise, arrayname is treated as in
the ARRAY commands: it can be a, &a, &a[], \&a, \&a[], \&a[3],
\&a[3:9], etc, with the appropriate results in each case.
Synonym: IF DCL.
UNDECLARE arrayname
UNDECLARE is a new top-level command to undeclare an array.
Previously this could only be done with "declare \&a[0]" (i.e.
re-declare the array with a dimension of 0).
ARRAY LINK linkname arrayname
Creates a symbolic link from the array named by linkname (which
must be the name of an array that is not yet declared in the
current context) to the array named by arrayname (which must the
name of a currently declared array that is not itself a link, or
a variable containing the name of such an array). The two names
indicate the same array: if you change an array element, the
change is reflected in the link too, and vice versa. If you
undeclare the link, the real array is unaffected. If you
undeclare the real array, all links to it disappear. If you
resize an array (directly or through a link), all links to it
are updated automatically.
Array links let you pass array names as arguments to macros. For
example, suppose you had a program that needed to uppercase all the
elements of different arrays at different times. You could write a
macro to do this, with the array name as an argument. But without array
links, there would be no way to refer to the argument array within the
macro. Array links make it easy:
define arrayupper {
local \&e[] \%i
array link \&e[] \%1
for i 1 \fdim(&e) 1 { .\&e[i] := \fupper(\&e[i]) }
}
declare \&a[] = these are some words
arrayupper &a
show array &a
The macro declares the array link LOCAL, which means it doesn't
conflict with any array of the same name that might exist outside the
macro, and that the link is destroyed automatically when the macro
exits. This works, by the way, even if the link name and the macro
argument name are the same, as long as the link is declared LOCAL.
As noted, you can't make a link to a nonexistent array. So when writing
a macro whose job is to create an array whose name is passed as an
argument, you must declare the array first (the size doesn't matter as
long as it's greater than 0). Example:
define tryme { ; Demonstration macro
local \&e[] ; We only need this inside the macro
array link \&e[] \%1 ; Make local link
shift ; Shift argument list
void \fsplit(\%*,&e) ; Split remainder of arg list into array
}
declare \&a[1] ; Declare target array in advance
tryme &a here are some words ; Invoke the macro with array name and words
show array a ; See the results
One final improvement allows the macro itself to declare the array
(this was not possible in earlier Kermit releases): if the array name
in the DECLARE command is a variable (and not an array name), or
includes variables, the resulting value is used as the array name. So:
define tryme { ; Demonstration macro
declare \%1[1] ; Preliminary declaration for target array
local \&e[] ; We only need this inside the macro
array link \&e[] \%1 ; Make local link
shift ; Shift argument list
void \fsplit(\%*,&e) ; Split remainder of arg list into array
}
tryme &a here are some words ; Invoke the macro with array name and words
show array a ; See the results
The SHOW ARRAY command now indicates whether an array name is a link.
Also see the descriptions of [428]\fjoin() and [429]\fsplit(), plus
[430]Section 8.10 on the MINPUT command, which shows how an entire
array (or segment of it) can be used as the MINPUT target list.
[ [431]Top ] [ [432]Contents ] [ [433]C-Kermit Home ] [ [434]Kermit
Home ]
8.7. New or Improved Built-in Variables and Functions
The following new built-in variables are available:
\v(buildid) A date string like "20000808" indicating when C-Kermit was b
uilt.
\v(ftime) Current time, secs since midnight, including fraction of sec
ond.
\v(iprompt) The current SET PROMPT value
\v(sexp) The most recent S-Expression (see [435]Section 9)
\v(sdepth) The current S-Expression invocation depth ([436]Section 9)
\v(svalue) The value of the most recent S-Expression ([437]Section 9)
\v(ftp_code) Most recent FTP response code ([438]Section 3)
\v(ftp_connected) FTP connection status ([439]Section 3)
\v(ftp_cpl) FTP Command Protection Level ([440]Section 3.2)
\v(ftp_dpl) FTP Data Protection Level ([441]Section 3.2)
\v(ftp_getputremote) The current SET GET-PUT-REMOTE setting ([442]Section 3.8)
\v(ftp_host) Name or IP address of FTP server ([443]Section 3)
\v(ftp_loggedin) FTP login status ([444]Section 3)
\v(ftp_message) Most recent FTP response message ([445]Section 3)
\v(ftp_security) FTP Security method ([446]Section 3.2)
\v(ftp_server) OS type of FTP server ([447]Section 3)
\v(http_code) Most recent HTTP response code
\v(http_connected) HTTP connection status
\v(http_host) Name or IP address of HTTP server
\v(http_message) Most recent HTTP response message
\v(http_security) TLS cipher used to secure the HTTP session
\v(hour) Hour of the day, 0 to 23.
\v(timestamp) Equivalent to "\v(ndate) \v(time)".
\v(log_debug) Current debug log file, if any.
\v(log_packet) Current packet log file, if any.
\v(log_session) Current session log file, if any.
\v(log_transaction) Current transaction log file, if any.
\v(log_connection) Current connection log file, if any.
The following new or improved built-in functions are available:
\fcmdstack() Allows programmatic access to the command stack.
\fcvtdate() [448]Section 8.13, format options added
\fdelta2secs() [449]Section 8.13
\fdostounixpath(s1) Converts a DOS filename to Unix format.
\fsplit() Now allows grouping/nesting in source string.
\fword() Allows the same grouping and nesting.
\fjoin(&a,s1,n1,n2) Copies an array into a single string.
\fsubstitute(s1,s2,s3) Substitutes characters within a string.
\freplace() Has new 4th "occurrence" argument.
\fsexpression() Evaluates an S-Expression (explained in [450]Section 9
).
\ftrim(), \fltrim() Now trim CR and LF by default, as well as SP and Tab.
\funixtodospath(s1) Converts a Unix filename to DOS format.
\fkeywordval(s1,c1) Assigns values to keywords (macros) (explained below).
Most functions that have "2" in their names to stand for the word "to"
can now also be written with "to", e.g. "\fdelta2secs(),"
\fdeltatosecs()."
\funtabify(string)
(New to 8.0.211) Replaces Horizontal Tab characters in the given
string with spaces based on VT100-like tab stops.
\fverify(s1,s2,n)
As of version 8.0.211, returns -1 if s2 is an empty string.
Previously it returned 0, making \fverify(abc,\%a) look as if
\%a was a string composed of a's, b's, and/or c's when in fact
it contained nothing.
\fcode(string)
As of version 8.0.211, returns 0 if string is empty or missing.
Previously it returned the empty string, which made it unsafe to
use in arithmetic or boolean expressions.
\v(inscale)
New to version 8.0.211, its value is the INPUT SCALE-FACTOR
([451]Section 8.10), default 1.0.
8.7.1. The \fkeywordval() Function
\fkeywordval(s1,c1) is new to C-Kermit 8.0. Given a string s1 of the
form "name=value", it creates a macro with the given name and assigns
it the given value. If no value appears after the equal sign, any
existing macro of the given name is undefined. Blanks are automatically
trimmed from around the name and value. The optional c1 parameter is
the assignment operator character, equal sign (=) by default. This
function is handy for processing keyword parameters or any other form
of parameter-value pair. Suppose, for example, you want to write a
macro that accepts keyword parameters rather than positional ones:
define MYDIAL {
local \%i modem hangup method device speed number
def number 5551234 ; Assign default parameter values
def speed 57600
def modem usrobotics
def hangup rs232
def method tone
def country 1
for \%i 1 \v(argc)-1 1 { ; Parse any keyword parameters...
if not \fkeywordval(\&_[\%i]) end 1 Bad parameter: "\&_[\%i]"
}
set dial country \m(country)
set modem type \m(modem)
set modem hang \m(hangup)
set dial method \m(tone)
set line \m(device)
if fail stop 1
set speed \m(speed)
if fail stop 1
show comm
set dial display on
dial \m(number)
if success connect
}
In this example, all the defaults are set up inside the macro, and
therefore it can be invoked with no parameters at all. But if you want
to have the macro dial a different number, you can supply it as
follows:
mydial number=7654321
You can supply any number of keyword parameters, and you can give them
in any order:
mydial number=7654321 hangup=modem speed=115200
8.7.2. The \fsplit(), \fjoin(), and \fword() Functions
\fjoin(&a,s1,n1,n2) is also new; it creates a string from an array (or
a piece of one). &a is the name of the array (a range specifier can be
included); s1 is a character or string to separate each element in the
result string (can be omitted, in which case the elements are not
separated at all), and n1 is a grouping mask, explained below. If s1 is
empty or not specified, the array elements are separated with spaces.
If you want the elements concatenated with no separator, include a
nonzero n2 argument. Given the array:
declare \&a[] = 0 1 2 3 4 5 6 7 8 9
you can get effects like this:
\fjoin(&a) 0 1 2 3 4 5 6 7 8 9
\fjoin(&a,:) 0:1:2:3:4:5:6:7:8:9
\fjoin(&a,{,}) 0,1,2,3,4,5,6,7,8,9
\fjoin(&a,...) 0...1...2...3...4...5...6...7...8...9
\fjoin(&a,,,1) 0123456789
\fsplit(), \fword(), \fstripb(), and \fjoin() accept a "grouping mask"
argument, n1, which is a number from 0 to 63, in which:
1 = "" doublequotes
2 = {} braces
4 = '' singlequotes
8 = () parentheses
16 = [] square brackets
32 = <> angle brackets
These can be OR'd (added) together to make any number 0-63 (-1 is
treated the same as 63, 0 means no grouping). If a bit is on, the
corresponding kind of grouping is selected. (If more than 1 bit is set
for \fjoin(), only the lowest-order one is used.)
If you include the same character in the grouping mask and the include
list, the grouping mask takes precedence. Example:
def \%a a "b c d" e
\fsplit(\%a,&a[],,,-1) = 3 <-- doublequote used for grouping
\fsplit(\%a,&a[],,",-1) = 3 <-- doublequote still used for grouping
Nesting of matched left and right grouping characters (parentheses,
braces, and brackets, but not quotes) is recognized. Example:
def \%a a (b c <d e [f g {h i} j k] l m> n o) p
\fsplit(\%a,&a,,,0) = 16 (no grouping)
\fsplit(\%a,&a,,,2) = 15 (braces only)
\fsplit(\%a,&a,,,16) = 11 (square brackets only)
\fsplit(\%a,&a,,,32) = 7 (angle brackets only)
\fsplit(\%a,&a,,,63) = 3 (all)
\fsplit(\%a,&a,,,-1) = 3 (all)
\fsplit() and \fjoin() are "reciprocal" functions. You can split a
string up into an array and join it back into a new string that is
equivalent, as long as \fsplit() and \fjoin() are given equivalent
grouping masks, except that the type of braces might change. Example:
def \%a a {b c [d e] f g} "h i" j <k l> m
echo STRING=[\%a]
echo WORDS=\fsplit(\%a,&a,,,-1)
show array a
asg \%b \fjoin(&a,{ },2)
echo JOIN =[\%b]
echo WORDS=\fsplit(\%b,&b,,,-1)
show array b
The arrays a and b are identical. The strings a and b are as follows:
\%a: a {b c [d e] f g} "h i" j <k l> m
\%b: a {b c [d e] f g} {h i} j {k l} m
It is possible to quote separator grouping characters with backslash to
override their grouping function. And of course to include backslash
itself in the string, it must be quoted too. Furthermore, each
backslash must be doubled, so the command parser will still pass one
backslash to \fsplit() for each two that it sees. Here are some
examples using \fsplit() with a grouping mask of 8 (treat parentheses
as grouping characters).
String Result
a b c d e f 6
a b\\ c d e f 5
a b (c d e) f 4
a b \\(c d e\\) f 6
a b \\\\(c d e\\\\) f 7
\fsplit() has also been changed to create its array (if one is given)
each time it is called, so now it can be conveniently called in a loop
without having to redeclare the array each time.
Incidentally... Sometimes you might want to invoke \fsplit() in a
situation where you don't care about its return value, e.g. when you
just want to fill the array. Now you can "call" \fsplit() or any other
function with the new [452]VOID command:
void \fsplit(\%a,&a)
\fsplit() and \fjoin() also accept a new, optional 6th argument, an
options flag, a number that can specify a number of options. So far
there is just one option, whose value is 1:
separator-flag
Normally separators are collapsed. So, for example,
\fword(Three little words,2)
returns "little" (the second word). Space is a separator, but
there are multiple spaces between each word. If the value 1 is
included in the option flag, however, each separator counts. If
two separators are adjacent, an empty word is produced between
them. This is useful for parsing (e.g.) comma-separated lists
exported from databases or spreadsheets.
8.7.3. The \fcmdstack() Function
The new \fcmdstack() function gives access to the command stack:
\fcmdstack(n1,n2)
Arguments: n1 is the command stack level. If omitted, the
current level, \v(cmdlevel), is used. n2 is a function code
specifying the desired type of information:
0 (default) = name of object at level n1.
1 (nonzero) = object type (0 = prompt; 1 = command file; 2 = macro).
The default for n2 is 0.
The name associated with prompt is "(prompt)". Here's a loop that can
be included in a macro or command file to show the stack (similar to
what the SHOW STACK command does):
for \%i \v(cmdlevel) 0 -1 {
echo \%i. [\fcmdstack(\%i,1)] \fcmdstack(\%i,0)
}
In this connection, note that \v(cmdfile) always indicates the most
recently invoked active command file (if any), even if that file is
executing a macro. Similarly, \v(macro) indicates the most recently
invoked macro (if any), even if the current command source is not a
macro. The name of the "caller" of the currently executing object
(command file or macro) is:
\fcmdstack(\v(cmdlevel)-1)
and its type is:
\fcmdstack(\v(cmdlevel)-1,1)
To find the name of the macro that invoked the currently executing
object, even if one or more intermediate command files (or prompting
levels) are involved, use a loop like this:
for \%i \v(cmdlevel)-1 0 -1 {
if = \fcmdstack(\%i,1) 2 echo CALLER = \fcmdstack(\%i,0)
}
Of course if you make a macro to do this, the macro must account for
its own additional level:
define CALLER {
for \%i \v(cmdlevel)-2 0 -1 {
if = \fcmdstack(\%i,1) 2 return \fcmdstack(\%i,0)
}
return "(none)"
}
The built-in variable \v(cmdsource) gives the current command source as
a word ("prompt", "file", or "macro").
8.7.4. The VOID Command
VOID is like ECHO in that all functions and variables in its argument
text are evaluated. but it doesn't print anything (except possibly an
error message if a function was invocation contained or resulted in any
errors). VOID sets FAILURE if it encounters any errors, SUCCESS
otherwise.
[ [453]Top ] [ [454]Contents ] [ [455]C-Kermit Home ] [ [456]Kermit
Home ]
8.8. The RETURN and END Commands
The execution of a macro is terminated in any of the following ways:
* With an END [ number [ message ] ] command. If a number is given,
the macro succeeds if the number is 0, and fails if it is not zero;
if a number is not given, the macro succeeds.
* With a STOP command, which works just like END except it peels back
the command stack all the way to top level.
* With a RETURN [ text ] command, in which case the macro always
succeeds.
* By running out of commands to execute, in which case the macro
succeeds or fails according the most recently executed command that
sets success or failure.
The same considerations apply to command files invoked by the TAKE
command.
If a macro does not execute any commands that set success or failure,
then invoking the macro does not change the current SUCCESS/FAILURE
status. It follows, then, that the mere invocation of a macro does not
change the SUCCESS/FAILURE status either. This makes it possible to
write macros to react to the status of other commands (or macros), for
example:
define CHKLINE {
if success end 0
stop 1 SET LINE failed - please try another device.
}
set modem type usrobotics
set line /dev/cua0
chkline
set speed 57600
dial 7654321
By the way, none of this is news. But it was not explicitly documented
before, and C-Kermit 7.0 and earlier did not always handle the RETURN
statement as it should have.
[ [457]Top ] [ [458]Contents ] [ [459]C-Kermit Home ] [ [460]Kermit
Home ]
8.9. UNDEFINing Groups of Variables
The UNDEFINE command, which previously accepted one variable name, now
accepts a list of them, and also accepts wildcard notation to allow
deletion of variables that match a given pattern.
UNDEFINE [ switches ] name [ name [ name [ ... ] ] ]
Undefines the variables whose names are given. Up to 64 names
may be given in one UNDEFINE command.
If you omit the switches and include only one name, the UNDEFINE
command works as before.
Switches include:
/MATCHING
Specifies that the names given are to treated as patterns rather
than literal variable names. Note: pattern matching can't be
used with array references; use the ARRAY command to manipulate
arrays and subarrays.
/LIST
List the name of each variable to be undefined, and whether it
was undefined successfully ("ok" or "error"), plus a summary
count at the end.
/SIMULATE
List the names of the variables that would be deleted without
actually deleting them. Implies /LIST.
The UNDEFINE command fails if there were any errors and succeeds
otherwise.
The new _UNDEFINE command is like UNDEFINE, except the names are
assumed to be variable names themselves, which contain the names (or
parts of them) of the variables to be undefined. For example, if you
have the following definitions:
define \%a foo
define foo This is some text
then:
undef \%a
undefines the variable \%a, but:
_undef \%a
undefines the macro foo.
Normal Kermit patterns are used for matching; metacharacters include
asterisk, question mark, braces, and square brackets. Thus, when using
the /MATCHING switch, if the names of the macros you want to undefine
contain any of these characters, you must quote them with backslash to
force them to be taken literally. Also note that \%* is not the name of
a variable; it is a special notation used within a macro for "all my
arguments". The command "undef /match \%*" deletes all \%x variables,
where x is 0..9 and a..z. Use "undef /match \%[0-9]" to delete macro
argument variables or "undef /match \%[i-n]" to delete a range of \%x
variables.
[ [461]Top ] [ [462]Contents ] [ [463]C-Kermit Home ] [ [464]Kermit
Home ]
8.10. The INPUT and MINPUT Commands
As of C-Kermit 8.0.211, the INPUT and MINPUT commands accept a switch:
[M]INPUT /NOMATCH timeout
The /NOMATCH switch allows INPUT or MINPUT to read incoming
material for the specified amount of time, without attempting to
match it with any text or patterns. When this switch is
included, the [M]INPUT command succeeds when the timeout
interval expires, with \v(instatus) set to 1, meaning "timed
out", or fails upon interruption or i/o error.
Also in version 8.0.211, there is a new way to apply a scale factor to
[M]INPUT timeouts:
SET INPUT SCALE-FACTOR floating-point-number
This scales all [M]INPUT timeouts by the given factor, allowing
time-sensitive scripts to be adjusted to changing conditions
such as congested networks or different-speed modems without
having to change each INPUT-class command. This affects only
those timeouts that are given in seconds, not as wall-clock
times. Although the scale factor can have a fractional part, the
INPUT timeout is still an integer. The new built-in variable
\v(inscale) tells the current INPUT SCALE-FACTOR.
The MINPUT command can be used to search the incoming data stream for
several targets simultaneously. For example:
MINPUT 8 one two three
waits up to 8 seconds for one of the words "one", "two", or "three" to
arrive. Words can be grouped to indicate targets that contain spaces:
MINPUT 8 nineteen twenty "twenty one"
And of course you can also use variables in place of (or as part of)
the target names:
MINPUT 8 \%a \&x[3] \m(foo)
Until now you had to know the number of targets in advance when writing
the MINPUT statement. Each of the examples above has exactly three
targets.
But suppose your script needs to look for a variable number of targets.
For this you can use arrays and \fjoin(), described in [465]Section
8.7. Any number of \fjoin() invocations can be included in the MINPUT
target list, and each one is expanded into the appropriate number of
separate targets each time the MINPUT command is executed. Example:
declare \&a[10] = one two three
minput 10 foo \fjoin(&a) bar
This declares an array of ten elements, and assigns values to the first
three of them. The MINPUT command looks for these three (as well as the
words "foo" and "bar"). Later, if you assign additional elements to the
array, the same MINPUT command also looks for the new elements.
If an array element contains spaces, each word becomes a separate
target. To create one target per array element, use \fjoin()'s grouping
feature:
dcl \&a[] = {aaa bbb} {ccc ddd} {xxx yyy zzz}
minput 10 \fjoin(&a) <-- 7 targets
minput 10 \fjoin(&a,,2) <-- 3 targets
[ [466]Top ] [ [467]Contents ] [ [468]C-Kermit Home ] [ [469]Kermit
Home ]
8.11. Learned Scripts
C-Kermit now includes a simple script recorder that monitors your
commands, plus your actions during CONNECT mode, and automatically
generates a script program that mimics what it observed. You should
think of this feature as a script-writing ASSISTANT since, as you will
see [470]later in this section, the result generally needs some editing
to make it both secure and flexible. The script recorder is controlled
by the new LEARN command:
LEARN [ /ON /OFF /CLOSE ] [ filename ]
If you give a filename, the file is opened for subsequent
recording. The /ON switch enables recording to the current file
(if any); /OFF disables recording. /CLOSE closes the current
script recording file (if any). If you give a filename without
any switches, /ON is assumed.
The /OFF and /ON switches let you turn recording off and on during a
session without closing the file.
When recording:
* All commands that you type (or recall) at the prompt are recorded
in the file except:
+ LEARN commands are not recorded.
+ The CONNECT command is not recorded.
+ The TELNET command is converted to SET HOST /NETWORK:TCP.
* Commands obtained from macros or command files are not recorded.
* During CONNECT:
+ Every line you type is converted to an OUTPUT command.
+ The last prompt before any line you type becomes an INPUT
command.
+ Timeouts are calculated automatically for each INPUT command.
+ A PAUSE command is inserted before each OUTPUT command just to
be safe.
Thus the script recorder is inherently line-oriented. It can't be used
to script character-oriented interactions like typing Space to a
"More?" prompt or editing a text file with VI or EMACS.
But it has advantages too; for example it takes control characters into
account that might not be visible to you otherwise, and it
automatically converts control characters in both the input and output
streams to the appropriate notation. It can tell, for example that the
"$ " prompt on the left margin in UNIX is really {\{13}\{10}$ },
whereas in VMS it might be {\{13}\{10}\{13}$ }. These sequences are
detected and recorded automatically.
A learned script should execute correctly when you give a TAKE command
for it. However, it is usually appropriate to edit the script a bit.
The most important change would be to remove any passwords from it. For
example, if the script contains:
INPUT 9 {\{13}\{10}Password: }
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT bigsecret\{13}
you should replace this by something like:
INPUT 9 {\{13}\{10}Password: }
IF FAIL STOP 1 INPUT timeout
ASKQ pswd Please type your password:
PAUSE 1
OUTPUT \m(pswd)\{13}
The LEARN command can't do this for you since it knows nothing about
"content"; it only knows about lines and can't be expected to parse or
understand them -- after all, the Password prompt might be in some
other language. So remember: if you use the LEARN command to record a
login script, be sure edit the resulting file to remove any passwords.
Also be sure to delete any backup copies your editor or OS might have
made of the file.
Other manual adjustments might also be appropriate:
* If the target of an INPUT command can vary, you can replace the
INPUT command with MINPUT and the appropriate target list, and/or
the target with a \fpattern(). For example, suppose you are dialing
a number that can be answered by any one of 100 terminal servers,
whose prompts are ts-00>, ts-01>, ts-02>, ... ts-99>. The script
records a particular one of these, but you want it to work for all
of them, so change (e.g.):
INPUT 10 ts-23> ; or whatever
to:
INPUT 10 \fpattern(ts-[0-9][0-9]>)
* The INPUT timeout values are conservative, but they are based only
on a single observation; you might need to tune them.
* The PAUSE commands might not be necessary, or the PAUSE interval
might need adjustment.
* In case you made typographical errors during recording, they are
incorporated in your script; you can edit them out if you want to.
Here is a sample script generated by Kermit ("learn vms.ksc") in which
a Telnet connection is made to a VMS computer, the user logs in, starts
Kermit on VMS, sends it a file, and then logs out:
; Scriptfile: vms.ksc
; Directory: /usr/olga
; Recorded: 20001124 15:21:23
SET HOST /NETWORK:TCP vms.xyzcorp.com
IF FAIL STOP 1 Connection failed
INPUT 7 {\{13}\{10}\{13}Username: }
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT olga\{13}
INPUT 3 {\{13}\{10}\{13}Password: }
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT secret\{13}
INPUT 18 {\{13}\{10}\{13}$ }
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT set default [.incoming]\{13}
INPUT 12 {\{13}\{10}\{13}$ }
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT kermit\{13}
INPUT 15 {\{13}\{10}\{13}ALTO:[OLGA.INCOMING] C-Kermit>}
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT receive\{13}
send myfile.txt
INPUT 18 {\{13}\{10}\{13}ALTO:[OLGA.INCOMING] C-Kermit>}
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT exit\{13}
INPUT 6 {\{13}\{10}\{13}$ }
IF FAIL STOP 1 INPUT timeout
PAUSE 1
OUTPUT logout\{13}
close
exit
The commands generated by Kermit during CONNECT (INPUT, IF FAIL, PAUSE,
and OUTPUT) have uppercase keywords; the commands typed by the user are
in whatever form the user typed them (in this case, lowercase).
[ [471]Top ] [ [472]Contents ] [ [473]C-Kermit Home ] [ [474]Kermit
Home ]
8.12. Pattern Matching
A pattern is a character string that is used to match other strings.
Patterns can contain metacharacters that represent special actions like
"match any single character", "match zero or more characters", "match
any single character from a list", and so on. The best known
application of patterns is in file specifications that contain
wildcards, as in "send *.txt", meaning "send all files whose names end
with .txt".
Patterns are also used in increasingly many other ways, to the extent
it is useful to point out certain important distinctions in the ways in
which they are used:
Anchored Patterns
If an anchored pattern does not begin with "*", it must match
the beginning of the string, and if it does not end with "*", it
must match the end of the string. For example, the anchored
pattern "abc" matches only the string "abc", not "abcde" or
"xyzabc" or "abcabc". The anchored pattern "abc*" matches any
string that starts with "abc"; the anchored pattern "*abc"
matches any string that ends with "abc"; the anchored pattern
"*abc*" matches any string that contains "abc" (including any
that start and/or end with it).
Floating Patterns
A floating pattern matches any string that contains a substring
that matches the pattern. In other words, a floating pattern has
an implied "*" at the beginning and end. You can anchor a
floating pattern to the beginning by starting it with "^", and
you can anchor it to the end by ending it with "$" (see examples
below).
Wildcards
A wildcard is an anchored pattern that has the additional
property that "*" does not match directory separators.
This terminology lets us describe Kermit's commands with a bit more
precision. When a pattern is used for matching filenames, it is a
wildcard, except in the TEXT-PATTERNS and BINARY-PATTERNS lists and
/EXCEPT: clauses, in which case directory separators are not
significant (for example, a BINARY-PATTERN of "*.exe" matches any file
whose name ends in .exe, no matter how deeply it might be buried in
subdirectories). When Kermit parses a file specification directly,
however, it uses the strict wildcard definition. For example, "send
a*b" sends all files whose names start with "a" and end with "b" in the
current directory, and not any files whose names end with "b" that
happen to be in subdirectories whose names start with "a". And as
noted, wildcards are anchored, so "delete foo" deletes the file named
"foo", and not all files whose names happen to contain "foo".
Most other patterns are anchored. For example:
if match abc bc ...
does not succeed (and you would be surprised if it did!). In fact, the
only floating patterns are the ones used by commands or functions that
search for patterns in files, arrays, or strings. These include:
* The GREP and TYPE /MATCH commands.
* The \fsearch(), \frsearch(), and \farraylook() functions.
Thus these are the only contexts in which explicit anchors ("^" and
"$") may be used:
grep abc *.txt
Prints all lines containing "abc" in all files whose names end
with ".txt".
grep ^abc *.txt
Prints all lines that start with "abc" in all ".txt" files.
grep abc$ *.txt
Prints all lines that end with "abc" in all ".txt" files.
grep ^a*z$ *.txt
Prints all lines that start with "a" and end with "z" in all
".txt" files.
Similarly for TYPE /PAGE, /fsearch(), /frsearch(), and \farraylook().
Here is a brief summary of anchored and floating pattern equivalences:
Anchored Floating
abc ^abc$
*abc abc$
abc* ^abc
*abc* abc
[ [475]Top ] [ [476]Contents ] [ [477]C-Kermit Home ] [ [478]Kermit
Home ]
8.13. Dates and Times
C-Kermit's comprehension of date-time formats is considerably expanded
in version 8.0. Any command that reads dates, including the DATE
command itself, or any switch, such as the /BEFORE: and /AFTER:
switches, or any function such as \fcvtdate(), now can understand dates
and times expressed in any ISO 8601 format, in Unix "asctime" format,
in FTP MDTM format, and in practically any format used in RFC 822 or
RFC 2822 electronic mail, with or without timezones, and in a great
many other formats as well. HELP DATE briefly summarizes the acceptable
date-time formats.
Furthermore, C-Kermit 8.0 includes a new and easy-to-use form of
date-time arithmetic, in which any date or time can be combined with a
"delta time", to add or subtract the desired time interval (years,
months, weeks, days, hours, minutes, seconds) to/from the given date.
And new functions are available to compare dates and to compute their
differences.
As you can imagine, all this requires quite a bit of "syntax". The
basic format is:
[ date ] [ time ] [ delta ]
Each field is optional, but in most cases (depending on the context)
there must be at least one field. If a date is given, it must come
first. If no date is given, the current date is assumed. If no time is
given, an appropriate time is supplied depending on whether a date was
supplied. If no delta is given, no arithmetic is done. If a delta is
given without a date or time, the current date and time are used as the
base.
Date-time-delta fields are likely to contain spaces (although they need
not; space-free forms are always available). Therefore, in most
contexts -- and notably as switch arguments -- date-time information
must be enclosed in braces or doublequotes, for example:
send /after:"8-Aug-2001 12:00 UTC" *.txt
Kermit's standard internal format for dates and times is:
yyyymmdd hh:mm:ss
for example:
20010208 10:28:01
Date-times can always be given in this format. yyyy is the 4-digit
year, mm is the two-digit month (1-12; supply leading zero for
Jan-Sep), dd is the 2-digit day (leading zero for 1-9), hh is the hour
(0-23), mm the minute (0-59), ss the second (0-59), each with leading
zero if less than the field width. The date and time can be separated
by a space, an underscore, a colon, or the letter T. The time is in
24-hour format. Thus the various quantities are at the following fixed
positions:
Position Contents
1-4 Year (4 digits, 0000-9999)
5-6 Month (2 digits, 1-12)
7-8 Day (2 digits, 1-31)
9 Date-Time Separator (space, :, _, or the letter T)
10-11 Hour (2 digits, 0-23)
12 Hour-Minute Separator (colon)
13-14 Minute (2 digits, 0-59)
15 Minute-Second Separator (colon)
16-17 Second (2 digits, 0-59)
Example:
19800526 13:07:12 26 May 1980, 13:07:12 (1:07:12PM)
This is the format produced by the DATE command and by any function
that returns a date-time. It is suitable for lexical comparison and
sorting, and for use as a date-time in any Kermit command. When this
format is given as input to a command or function, various date-time
separators (as noted) are accepted:
19800526 13:07:12 26 May 1980, 13:07:12 (1:07:12PM)
20010208_10:28:35 2 February 2001, 10:28:35 AM
18580101:12:00:00 1 January 1858, noon
20110208T00:00:00 2 February 2011, midnight
Certain other special date-time formats that are encountered on
computer networks are recognized:
Asctime Format
This is a fixed format used by Unix, named after Unix's
asctime() ("ASCII time") function. It is always exactly 24
characters long. Example: Fri Aug 10 16:38:01 2001
Asctime with Timezone
This is like Asctime format, but includes a 3-character timezone
between the time and year. It is exactly 28 characters long.
Example: Fri Aug 10 16:38:01 GMT 2001
E-Mail Format
E-mail date-time formats are defined in [479]RFC 2822 with a
fair amount of flexibility and options. The following examples
are typical of e-mails and HTTP (web-page) headers:
Sat, 14 Jul 2001 11:49:29 (No timezone)
Fri, 24 Mar 2000 14:19:59 EST (Symbolic timezone)
Tue, 26 Jun 2001 10:19:45 -0400 (EDT) (GMT Offset + comment)
FTP MDTM Format
This is the date-time format supplied by FTP servers that
support the (not yet standard but widely used nevertheless) MDTM
command, by which the FTP client asks for a file's modification
time:
yyyymmddhhmmss[.ffff]
where yyyy is the 4-digit year, mm is the 2-digit month, and so
on, exactly 14 digits long. An optional fractional part
(fraction of second) may also be included, separated by a
decimal point (period). Kermit rounds to the nearest second.
Example:
20020208102835.515 (8 February 2002 10:28:36 AM)
8.13.1. The Date
The date, if given, must precede the time and/or delta, and can be in
many, many formats. For starters, you can use several symbolic date
names in place of actual dates:
NOW
This is replaced by the current date and time. The time can not
be overridden (if you want to supply a specific time, use TODAY
rather than NOW).
TODAY
This is replaced by the current date and a default time of
00:00:00 is supplied, but can be overridden by a specific time;
for example, if today is 8 February 2002, then "TODAY" is
"20020802 00:00:00" but "TODAY 10:28" is "20020802 10:28:00".
TOMORROW
Like TODAY, but one day later (if today is 8 February 2002, then
"TOMORROW" is "20020803 00:00:00" but "TOMORROW 16:30" is
"20020803 16:30:00").
YESTERDAY
Like TODAY, but one day earlier.
MONDAY, TUESDAY, WEDNESDAY, ..., SUNDAY
The date on the given day of the week, today or later. A default
time of 00:00:00 is supplied but can be overridden. Example:
"SATURDAY 12:00" means next Saturday (or today, if today is
Saturday) at noon.
You can give an explicit date in almost any conceivable format, but
there are some rules:
* If a date is given, it must have three fields: day, month, and
year; the order can vary (except that the month can not be last).
* If names are used for days, months, etc, they must be English.
* The year must lie between 0000 and 9999, inclusive.
* All calendar calculations use Gregorian dating, so calculated dates
for years prior to 1582 (or later, depending on the country) will
not agree with historical dates. Other forms of dating (e.g.
Hebrew, Chinese) are not supported.
Various date-field separators are accepted: hyphen, slash, space,
underscore, period. The same field separator (if any) must be used in
both places; for example 18-Sep-2001 but not 18-Sep/2001. Months can be
numeric (1-12) or English names or abbreviations. Month name
abbreviations are normally three letters, e.g. Apr, May, Jun, Jul.
Capitalization doesn't matter.
Here are a few examples:
18 Sep 2001 (English month, abbreviated)
18 September 2001 (English month, spelled out)
2001 Sept 18 (Year, month, day)
18-Sep-2001 (With hyphens)
18/09/2001 (All numeric with slashes)
18.09.2001 (Ditto, with periods)
18_09_2001 (Ditto, with underscores)
09/18/2001 (See below)
2001/09/18 (See below)
September 18, 2001 (Correspondence style)
Sep-18-2001 (Month-day-year)
20010918 (Numeric, no separators)
You can also include the day of the week with a specific date, in which
case it is accepted (if it is a valid day name), but not verified to
agree with the given date:
Tue, 18 Sep 2001 (Abbreviated, with comma)
Tue,18 Sep 2001 (Comma but no space)
Tue 18 Sep 2001 (Abbreviated, no comma)
Tuesday 18 Sep 2001 (Spelled out)
Tuesday, 18 Sep 2001 (etc)
Friday, 18 Sep 2001 (Accepted even if not Friday)
In all-numeric dates with the year last, such as 18/09/2001, Kermit
identifies the year because it's 4 digits, then decides which of the
other two numbers is the month or day based on its value. If both are
12 or less and are unequal, the date is ambiguous and is rejected. In
all-numeric dates with the year first, the second field is always the
month and the third is the day. The month never comes last. A date with
no separators is accepted only if it is all numeric and has exactly
eight digits, and is assumed to be in yyyymmdd format.
20010918 (18-Sep-2001 00:00:00)
or 14 digits (as in FTP MDTM format):
20010918123456 (18-Sep-2001 12:34:56)
You can always avoid ambiguity by putting the year first, or by using
an English, rather than numeric, month. A date such as 09/08/2001 would
be ambiguous but 2001/09/08 is not, nor is 09-Aug-2001.
Until the late 1990s, it was common to encounter 2-digit years, and
these are found to this day in old e-mails and other documents. Kermit
accepts these dates if they have English months, and interprets them
according to the windowing rules of [480]RFC 2822: "If a two digit year
is encountered whose value is between 00 and 49, the year is
interpreted by adding 2000, ending up with a value between 2000 and
2049. If a two digit year is encountered with a value between 50 and
99, or any three digit year is encountered, the year is interpreted by
adding 1900."
If you need to specify a year prior to 1000, use leading zeros to
ensure it is not misinterpreted as a "non-Y2K-compliant" modern year:
7-Oct-77 (19771007 00:00:00)
7-Oct-0077 (00771007 00:00:00)
8.13.2. The Time
The basic time format is hh:mm:dd; that is hours, minutes, seconds,
separated by colons, perhaps with an optional fractional second
separated by a decimal point (period). The hours are in 24-hour format;
12 is noon, 13 is 1pm, and so on. Fields omitted from the right default
to zero. Fields can be omitted from the left or middle by including the
field's terminating colon. Examples:
11:59:59 (11:59:59 AM)
11:59 (11:59:00 AM)
11 (11:00:00 AM)
11:59:59.33 (11:59:59 AM)
11:59:59.66 (Noon)
03:21:00 (3:21:00 AM)
3:21:00 (3:21:00 AM)
15:21:00 (3:21:00 PM)
:21:00 (00:21:00 AM)
::01 (00:00:01 AM)
11::59 (11:00:59 AM)
Leading zeros can be omitted, but it is customary and more readable to
keep them in the minute and second fields:
03:02:01 (03:02:01 AM)
3:02:01 (03:02:01 AM)
3:2:1 (03:02:01 AM)
AM/PM notation is accepted if you wish to use it:
11:59:59 (11:59:59 AM)
11:59:59AM (11:59:59 AM)
11:59:59A.M. (11:59:59 AM)
11:59:59am (11:59:59 AM)
11:59:59a.m. (11:59:59 AM)
11:59:59PM (11:59:59 PM = 23:59:59)
11:59:59P.M. (11:59:59 PM = 23:59:59)
11:59:59pm (11:59:59 PM = 23:59:59)
11:59:59p.m. (11:59:59 PM = 23:59:59)
You can omit the colons if you wish, in which case Kermit uses the
following rules to interpret the time:
1. 6 digits is hh:mm:ss, e.g. 123456 is 12:34:56.
2. 5 digits is h:mm:ss, e.g. 12345 is 1:23:45.
3. 4 digits is hh:mm, e.g. 1234 is 12:34.
4. 3 digits is h:mm, e.g. 123 is 1:23.
5. 2 digits is hh, e.g. 12 is 12:00.
6. 1 digit is h (the hour), e.g. 1 is 1:00.
Examples:
1 (01:00:00 AM)
10 (10:00:00 AM)
230 (02:30:00 AM)
230pm (02:30:00 PM = 14:30:00)
1115 (11:15:00 AM)
2315 (11:15:00 PM = 23:15:00 PM)
23150 (02:31:50 AM)
231500 (23:15:00 PM)
8.13.3. Time Zones
If a time is given, it can (but need not) be followed by a time zone
designator. If no time zone is included, the time is treated as local
time and no timezone conversions are performed.
The preferred time zone designator is the UTC Offset, as specified in
[481]RFC 2822: a plus sign or minus sign immediately followed by
exactly four decimal digits, signifying the difference in hh (hours)
and mm (minutes) from Universal Coordinated Time (UTC, also known as
Greenwich Mean Time, or GMT), with negative numbers to the West and
positive numbers to the East. For example:
Fri, 13 Jul 2001 12:54:29 -0700
indicates a local time of 12:54:29 that is 07 hours and 00 minutes
behind (less than, East of) Universal Time. The space is optional, so
the example could also be written as:
Fri, 13 Jul 2001 12:54:29-0700
The following symbolic time zones are also accepted, as specified by
[482]RFC 2822 and/or in ISO 8601:
GMT = +0000 Greenwich Mean Time
Z = +0000 Zulu (Zero Meridian) Time
UTC = +0000 Universal Coordinated Time
UT = +0000 Universal Time
EDT = -0400 Eastern (USA) Daylight Time
EST = -0500 Eastern (USA) Standard Time
CDT = -0500 Central (USA) Daylight Time
CST = -0600 Central (USA) Standard Time
MDT = -0600 Mountain (USA) Daylight Time
MST = -0700 Mountain (USA) Standard Time
PDT = -0700 Pacific (USA) Daylight Time
PST = -0800 Pacific (USA) Standard Time
Note that GMT, Z, UTC, and UT all express the same concept: standard
(not daylight) time at the Zero Meridian. UTC, by the way, is an
international standard symbol and does not correspond to the order of
the English words, Universal Coordinated Time, but it happens to have
the same initial letters as these words. Of course hundreds of other
symbolic timezones and variations exist, but they are not standardized,
and are therefore not supported by Kermit.
When a time zone is included with a time, the time is converted to
local time. In case the conversion crosses a midnight boundary, the
date is adjusted accordingly. Examples converting to EST (Eastern USA
Standard Time = -0500):
11:30:00 = 11:30:00
11:30:00 EST = 11:30:00
11:30:00 GMT = 06:30:00
11:30:00 PST = 14:30:00
11:30:00Z = 06:30:00
11:30PM GMT = 18:30:00
11:30 -0500 = 11:30:00
11:30 -0800 = 08:30:00
11:30 +0200 = 04:30:00
Unlike most of Kermit's other date-time conversions, timezone knowledge
(specifically, the offset of local time from UTC) is embodied in the
underlying operating system, not in Kermit itself, and any conversion
errors in this department are the fault of the OS. For example, most
UNIX platforms do not perform conversions for years prior to 1970.
8.13.4. Delta Time
Date/time expressions can be composed of a date and/or time and a delta
time, or a delta time by itself. When a delta time is given by itself,
it is relative to the current local date and time. Delta times have the
following general format:
{+,-}[number units][hh[:mm[:ss]]]
In other words, a delta time always starts with a plus or minus sign,
which is followed by a "part1", a "part2", or both. The "part1", if
given, specifies a number of days, weeks, months, or years; "part2"
specifies a time in hh:mm:ss notation. In arithmetic terms, these
represents some number of days or other big time units, and then a
fraction of a day expressed as hours, minutes, and seconds; these are
to be added to or subtracted from the given (or implied) date and time.
The syntax is somewhat flexible, as shown by the following examples:
+1 day (Plus one day)
+1day (Ditto)
+1d (Ditto)
+ 1 day (Ditto)
+ 1 day 3:00 (Plus one day and 3 hours)
+1d3:00 (Ditto)
+1d3 (Ditto)
+3:00:00 (Plus 3 hours)
+3:00 (Ditto)
+3 (Ditto)
+2 days (Plus 2 days)
-12 days 7:14:22 (Minus 12 days, 7 hours, 14 minutes, and 22 seconds)
The words "week", "month", and "year" can be used like "day" in the
examples above. A week is exactly equivalent to 7 days. When months are
specified, the numeric month number of the date is incremented or
decremented by the given number, and the year and day adjusted
accordingly if necessary (for example, 31-Jan-2001 +1month =
03-Mar-2001 because February does not have 31 days). When years are
specified, they are added or subtracted to the base year. Examples
(assuming the current date is 10-Aug-2001 and the current time is
19:21:11):
18-Sep-2001 +1day (20010918 00:00:00)
today +1day (20010811 00:00:00)
now+1d (20010811 19:21:11)
+ 1 day (20010811 19:21:11)
+ 1 day 3:14:42 (20010811 22:35:54)
+ 7 weeks (20010928 19:21:11)
+1d3:14:42 (20010811 22:35:54)
+1w3:14:42 (20010817 22:35:54)
+1m3:14:42 (20010910 22:35:54)
+1y3:14:42 (20020810 22:35:54)
2 feb 2001 + 10 years (20110208 00:00:00)
2001-02-08 +10y12 (20110208 12:00:00)
31-dec-1999 23:59:59+00:00:01 (20000101 00:00:00)
28-feb-1996 +1day (19960229 00:00:00) (leap year)
28-feb-1997 +1day (19970301 00:00:00) (nonleap year)
28-feb-1997 +1month (19970328 00:00:00)
28-feb-1997 +1month 11:59:59 (19970328 11:59:59)
28-feb-1997 +20years (20170228 00:00:00)
28-feb-1997 +8000years (99970228 00:00:00)
For compatibility with VMS, the following special delta-time format is
also accepted:
+number-hh:mm:ss
-number-hh:mm:ss
(no spaces). The hyphen after the number indicates days. It corresponds
exactly to the Kermit notation:
+numberdhh:mm:ss
-numberdhh:mm:ss
The following forms all indicate exactly the same date and time:
18-Sep-2001 12:34:56 +1-3:23:01
18-Sep-2001 12:34:56 +1d3:23:01
18-Sep-2001 12:34:56 +1 day 3:23:01
and mean "add a day plus 3 hours, 23 minutes, and 1 second" to the
given date.
Note that delta times are not at all the same as UTC offsets; the
former specifies an adjustment to the given date/time and the latter
specifies that the local time is a particular distance from Universal
Time, for example:
11-Aug-2001 12:34:56 -0800 (20010811 16:34:56 -- UTC Offset)
11-Aug-2001 12:34:56 -08:00 (20010811 04:34:56 -- Delta time)
If you give a time followed by a modifier that starts with a + or -
sign, how does Kermit know whether it's a UTC offset or a delta time?
It is treated as a UTC offset if the sign is followed by exactly four
decimal digits; otherwise it is a delta time. Examples (for USA Eastern
Daylight Time):
11-Aug-2001 12:34:56 -0800 (20010811 16:34:56 -- UTC Offset)
11-Aug-2001 12:34:56 -08:00 (20010811 04:34:56 -- Delta time)
11-Aug-2001 12:34:56 -800 (20010811 04:34:56 -- Delta time)
11-Aug-2001 12:34:56 -8 (20010811 04:34:56 -- Delta time)
The first example says that at some unknown place which is 8 hours
ahead of Universal Time, the time is 12:34:56, and this corresponds to
16:34:56 in Eastern Daylight time. The second example says to subtract
8 hours from the local time. The third and fourth are delta times
because, even though a colon is not included, the time does not consist
of exactly 4 digits.
When a delta time is written after a timezone, however, there is no
ambiguity and no syntax distinction is required:
11-Aug-2001 12:34:56 -0800 -0800 (20010811 08:34:56)
11-Aug-2001 12:34:56 -0800 -08:00 (Ditto)
11-Aug-2001 12:34:56 -08:00 -08:00 (Illegal)
8.13.5. The DATE Command
Obviously a great many combinations of date, time, time zone, and delta
time are possible, as well as many formatting options. The purpose of
all this flexibility is to comply with as many standards as possible --
Internet RFCs, ISO standards, and proven corporate standards -- as well
as with notations commonly used by real people, in order that dates and
times from the widest variety of sources can be assigned to a variable
and used in any date-time field in any Kermit command.
You can test any date-and/or-time format with the DATE command, which
converts it to standard yyyymmdd hh:mm:ss format if it is understood,
or else gives an explicit error message (rather than just "BAD DATE" as
in previous C-Kermit releases) to indicate what is wrong with it.
Examples (on Tuesday, 31 July 2001 in New York City, Eastern Daylight
Time, UTC -0400):
DATE command argument Result
12:30 20010731 12:30:00
12:30:01 20010731 12:30:01
12:30:01.5 20010731 12:30:02
1230 20010731 12:30:00
230 20010731 02:30:00
230+1d 20010801 02:30:00
230+1d3:00 20010801 05:30:00
20010718 19:21:15 20010718 19:21:15
20010718_192115 20010718 19:21:15
20010718T192115 20010718 19:21:15
18 Jul 2001 +0400 20010717 23:59:59
18 Jul 2001 192115 20010718 19:21:15
18 Jul 2001 192115.8 20010718 19:21:16
18-Jul-2001T1921 20010718 19:21:00
18-Jul-2001 1921Z 20010718 15:21:00
18-Jul-2001 1921 GMT 20010718 15:21:00
18-Jul-2001 1921 UTC 20010718 15:21:00
18-Jul-2001 1921 Z 20010718 15:21:00
18-Jul-2001 1921Z 20010718 15:21:00
18-Jul-2001 1921 -04:00:00 20010718 19:21:00
21-Jul-2001_08:20:00am 20010721 08:20:00
21-Jul-2001_8:20:00P.M. 20010721 20:20:00
Fri Jul 20 11:26:25 2001 20010720 11:26:25
Fri Jul 20 11:26:25 GMT 2001 20010720 07:26:25
Sun, 9 Apr 2000 06:46:46 +0100 20000409 01:46:46
Sunday, 9 Apr 2000 06:46:46 +0100 20000409 01:46:46
now 20010731 19:41:12
today 20010731 00:00:00
today 09:00 20010731 09:00:00
tomorrow 20010801 00:00:00
tomorrow 09:00 20010801 09:00:00
tomorrow 09:00 GMT 20010801 05:00:00
yesterday 20010730 00:00:00
yesterday 09:00 20010730 09:00:00
+ 3 days 20010803 00:00:00
+3 days 20010803 00:00:00
+3days 20010803 00:00:00
+ 3days 20010803 00:00:00
+ 3 days 09:00 20010803 09:00:00
+ 2 weeks 20010814 00:00:00
+ 1 month 20010831 00:00:00
- 7 months 20001231 00:00:00
+ 10 years 20110731 00:00:00
friday 20010803 00:00:00
saturday 20010804 00:00:00
sunday 20010805 00:00:00
monday 20010806 00:00:00
tuesday 20010731 00:00:00
wednesday 20010801 00:00:00
thursday 20010802 00:00:00
friday 07:00 20010803 07:00:00
thursday 1:00pm 20010802 13:00:00
thursday 1:00pm GMT 20010802 09:00:00
Thu, 10 Nov 94 10:50:47 EST 19941110 10:50:47
Fri, 20 Oct 1995 18:35:15 -0400 (EDT) 19951020 18:35:15
31/12/2001 20011231 00:00:00
12/31/2001 20011231 00:00:00
2001-July-20 20010720 00:00:00
2001-September-30 20010930 00:00:00
30-September-2001 20010930 00:00:00
Sep 30, 2001 12:34:56 20010930 12:34:56
September 30, 2001 20010930 00:00:00
September 30, 2001 630 20010930 06:30:00
September 30 2001 630 20010930 06:30:00
Sep-30-2001 12:34:59 20010930 12:34:59
20010807113542.014 20010807 11:35.42
20010807113542.014Z 20010807 07:35:42
8.13.6. New Date-Time Functions
In the following descriptions, date-time function arguments are the
same free-format date-time strings discussed above, with the same
defaults for missing fields. They are automatically converted to
standard format internally prior to processing.
\fcvtdate(d1)
Converts the date-time d1 to standard format and local time.
This function is not new, but now it accepts a wider range of
argument formats that can include timezones and/or delta times.
If the first argument is omitted, the current date and time are
assumed. The optional second argument is a format code for the
result:
n1 = 1: yyyy-mmm-dd hh:mm:ss (mmm = English 3-letter month
abbreviation)
n1 = 2: dd-mmm-yyyy hh:mm:ss (ditto)
n1 = 3: yyyymmddhhmmss (all numeric)
\futcdate(d1)
Converts the date-time d1 to Universal Coordinated Time (UTC),
also known as GMT or Zulu or Zero-Meridian time. The default d1
is NOW. If d1 is a valid date-time, the UTC result is returned
in standard format, yyyymmdd hh:ss:mm.
\fcmpdates(d1,d2)
Compares two free-format date-times, d1 and d2, and, if both
arguments are valid, returns a number: -1 if d1 is earlier than
(before) d2; 0 if d1 is the same as d2; 1 if d1 is later than
(after) d2.
\fdiffdates(d1,d2)
Computes the difference between two free-format date-times, d1
and d2. If both arguments are valid, returns a delta time which
is negative if d1 is earlier than (before) d2 and positive
otherwise. If d1 and d2 are equal, the result is "+0:00".
Otherwise, the result consists of the number of days, hours,
minutes, and seconds that separate the two date-times. If the
number of days is zero, it is omitted. If the number of days is
nonzero but the hours, minutes, and seconds are all zero, the
time is omitted. if the seconds are zero, they are omitted.
\fdelta2secs(dt)
Converts a delta time to seconds. For example, "+1d00:00:01" to
86401. Valid delta times must start with a + or - sign. Days are
accepted as time units, but not years, months, or weeks. If the
result would overflow a computer long word (as would happen with
32-bit long words when the number of days is greater than
24854), the function fails.
HINT: Although Kermit has a number of built-in date and time variables,
it doesn't have a single one suitable for writing a timestamp. For this
you would normally use something like "\v(ndate) \v(time)". But
\fcvtdate() (with no arguments) is equivalent: it returns the current
date and time in yyyymmdd hh:mm:ss format, suitable for time stamping.
8.13.7. Date-Time Programming Examples
Here's a macro that converts any date-time to UTC, which you might use
if C-Kermit didn't already have a \futcdate() function:
define utcdate {
.local := \fcvtdate(\%*) ; 1.
.tmp := \fcvtdate(\m(local)UTC) ; 2.
.offset := \fdiffdate(\m(local),\m(tmp)) ; 3.
.utc := \fcvtdate(\m(local)\m(offset)) ; 4.
sho mac utc ; 5.
}
Brief explanation: Line 1 converts the macro argument, a free-format
date-time, to standard-format local time. Line 2 appends the "UTC"
timezone to the local time and converts the result to local time. In
other words, we take the same time as the local time, but pretend it's
UTC time, and convert it to local time. For example, if New York time
is 4 hours ahead of UTC, then 6:00pm New York time is 2:00pm UTC. Line
3 gets the difference of the two results (e.g. "+04:00"). Line 4
appends the difference (delta time) to the local time, and converts it
again, which adds (or subtracts) the UTC offset to the given time. Line
5 displays the result.
Here's a script that opens a web page, gets its headers into an array,
scans the array for the "Last-Modified:" header, and interprets it:
http open www.columbia.edu
if fail stop 1 HTTP OPEN failed
http /array:a head index.html /dev/null
if fail stop 1 HTTP GET failed
show array a
for \%i 1 \fdim(&a) 1 {
.\%x := \findex(:,\&a[\%i])
if not \%x continue
.tag := \fleft(\&a[\%i],\%x-1)
.val := \fltrim(\fsubstr(\&a[\%i],\%x+1))
if ( eq "\m(tag)" "Last-Modified" ) {
echo HTTP Date: \m(val)
.rdate := \fcvtdate(\m(val))
echo {Standard Date (local): \m(rdate)}
echo {Standard Date (UTC): \futcdate(\m(rdate))}
break
}
}
http close
The result:
HTTP Date: Mon, 13 Aug 2001 20:05:42 GMT
Standard Date (local): 20010813 16:05:42
Standard Date (UTC): 20010813 20:05:42
As you can see, Kermit had no trouble decoding the date-time-string
from the website, converting to local time, and converting back to UTC
with no conflicts or loss of information. If it had been in any other
known format, the result would have been the same.
Now suppose we want to download the web page only if it is newer than
our local copy. The \fdate(filename) function (which returns the
modification date-time of the given file) and the new \fcmpdates()
function make it easy. Insert the following just before the BREAK
statement:
if ( < 0 \fcmpdates(\m(rdate),\fdate(index.html)) ) {
echo GETTING index.html...
http get index.html index.html
if success echo HTTP GET OK
} else {
echo index.html: no update needed
}
http close
exit
This says, "if 0 is less than the comparison of the remote file date
and the local file date, get the remote file, otherwise skip it." And
it automatically reconciles the time-zone difference (if any).
It would be nice to be able to extend this script into a
general-purpose website updater, but unfortunately HTTP protocol
doesn't provide any mechanism for the client to ask the server for a
list of files, recursive or otherwise.
[ [483]Top ] [ [484]Contents ] [ [485]C-Kermit Home ] [ [486]Kermit
Home ]
8.14. Trapping Keyboard Interruption
Normally when you type Ctrl-C and Kermit is in command mode (as opposed
to CONNECT mode) with COMMAND INTERRUPTION ON (as it is unless you have
set it OFF), Kermit interrupts any command that is currently in
progress, and if a command file or macro is executing, rolls the
command stack back to top level, closing all open command files,
deactivating all macros, deallocating all local variables and arrays,
and leaving you at the command prompt.
Suppose, however, you want certain actions to occur when a script is
interrupted; for example, closing open files, writing log entries, or
displaying summary results. You can do this by defining a macro named
ON_CTRLC. When Ctrl-C is detected, and a macro with this name is
defined, Kermit executes it from the current command level, thus giving
it full access to the environment in which the interruption occurred,
including local variables and open files. Only when the ON_CTRLC macro
completes execution is the command stack rolled back to top level.
Once the ON_CTRLC macro is defined, it can be executed only once. This
is to prevent recursion if the user types Ctrl-C while the ON_CTRLC
macro is executing. If you type Ctrl-C while the Ctrl-C macro is
active, this does not start a new copy of ON_CTRLC; rather, it returns
to the top-level command prompt. After the ON_CTRLC macro returns, it
has been removed from the macro table so if you want to use it again or
install a different Ctrl-C trap, you must execute a new DEFINE ON_CTRLC
command. In any case, as always when you interrupt a script with
Ctrl-C, its completion status is FAILURE.
Normally the ON_CTRLC macro would be defined in the command file or
macro to which it applies, and should be declared LOCAL. This way, if
the command file or macro completes successfully without being
interrupted, the ON_CTRLC definition disappears automatically.
Otherwise the definition would still be valid and the macro would be
executed, probably out of context, the next time you typed Ctrl-C.
Here's a simple example of a command file that sets a Ctrl-C trap for
itself:
local on_ctrlc ; Make Ctrl-C trap local to this command file.
define on_ctrlc { ; Define the ON_CTRLC macro.
echo Interrupted at \v(time).
echo Iterations: \%n
}
xecho Type Ctrl-C to quit
for \%n 1 999 1 { ; Prints a dot every second until interrupted.
sleep 1
xecho .
}
echo Finished normally at \v(time) ; Get here only if not interrupted.
decrement \%n
echo Iterations: \%n
This prints a summary no matter whether it completes normally or is
interrupted from the keyboard. In both cases the trap is automatically
removed afterwards.
For an example of how to use ON_CTRLC to debug scripts, see
[487]Section 8.1.
[ [488]Top ] [ [489]Contents ] [ [490]C-Kermit Home ] [ [491]Kermit
Home ]
9. S-EXPRESSIONS
This section is primarily for those who want to write
calculation-intensive scripts, especially if they require
floating-point arithmetic, and/or for those who are familiar with the
LISP programming language.
Ever since C-Kermit version 5 was released in 1988, scripting has been
one of its major attractions, and arithmetic is a key part of it.
Versions 5 and 6 included integer arithmetic only, using traditional
algebraic notation, e.g.:
echo \fevaluate(3*(2+7)/2)
13
C-Kermit 7.0 added support for floating-point arithmetic, but only
through function calls:
echo \ffpdivide(\ffpmultiply(3.0,\ffpadd(2.0,7.0)),2.0)
13.5
C-Kermit 8.0 introduces a third form of arithmetic that treats integers
and floating-point numbers uniformly, is easier to read and write, and
executes very quickly:
(/ (* 3 (+ 2 7)) 2)
13.5
But first some background.
The Kermit command and scripting language differs from true programming
languages (such as C or Fortran) in many ways; one of the most
prominent differences is the way in which variables are distinguished
from constants. In a command language, words are taken literally; for
example, the Unix shell:
cat foo.bar
displays the file named foo.bar. Whereas in a programming language like
C, words are assumed to be variables:
s = foo.bar; /* Assigns the value of foo.bar to the variable s */
To make a programming language take words literally, you have to quote
or "escape" them:
s = "foo.bar"; /* Assigns a pointer to the string "foo.bar" to the variable s
*/
The opposite holds for command languages: to get them to treat a word
as a variable rather than a constant, you have to escape them. For
example, in the Unix shell:
foo=123 ; Assign value 123 to variable foo.
echo foo ; Prints "foo"
echo $foo ; Prints "123"
And in Kermit:
define foo 123 ; Assign value 123 to variable foo.
echo 123 ; This prints "123".
echo foo ; This prints "foo".
echo \m(foo) ; This prints "123".
In other words, character strings (such as "foo" above) are interpreted
as literal strings, rather than variable names, except in special
commands like DEFINE that deal specifically with variable names (or in
numeric contexts as explained in [492]Section 8.2). The special
"escape" character (dollar sign ($) for the shell, backslash (\) for
Kermit) indicates that a variable is to be replaced by its value.
The requirement to escape variable names in command languages normally
does not impose any special hardship, but can add a considerable
notational burden to arithmetic expressions, which are typically full
of variables. Especially in Kermit when floating point numbers are
involved, where you must use special \ffpxxx() functions, e.g.
"\ffpadd(\m(a),\m(b))" rather than the simple "+" operator to add two
floating-point numbers together, because the original arithmetic
handler doesn't support floating point (this might change in the
future). To illustrate, the general formula for the area of a triangle
is:
sqrt(s * (s - a) * (s - b) * (s - c))
where a, b, and c are the lengths of the triangle's three sides and:
s = (a + b + c) / 2
Except in special cases (e.g. a = 3, b = 4, c = 5), the result has a
fractional part so the computation must be done using floating-point
arithmetic. We can create a Kermit 7.0 function for this as follows:
def area {
local s t1 t2 t3
assign s \ffpdiv(\ffpadd(\ffpadd(\%1,\%2),\%3),2.0)
assign t1 \ffpsub(\m(s),\%1)
assign t2 \ffpsub(\m(s),\%2)
assign t3 \ffpsub(\m(s),\%3)
return \ffpsqrt(\ffpmul(\m(s),\ffpmul(\m(t1),\ffpmul(\m(t2),\m(t3)))))
}
But as you can see, this is rather cumbersome. Note, in particular,
that arithmetic functions like \ffpadd(), \ffpmul(), etc, take exactly
two operands (like their symbolic counterparts + and *), so obtaining
the product of three or more numbers (as we do in this case) is
awkward.
Using the alternative S-Expression notation, we can reduce this to a
form that is both easier to read and executes faster (the details are
explained later):
def newarea {
(let s (/ (+ \%1 \%2 \%3) 2.0))
(sqrt (* s (- s \%1) (- s \%2) (- s \%3)))
}
In both examples, the \%1..3 variables are the normal Kermit macro
arguments, referenced by the normal escaping mechanism. For increased
readability, we can also assign the macro arguments \%1, \%2, and \%3
to the letters a, b, and c corresponding to our formula:
def newarea {
(let a \%1 b \%2 c \%3)
(let s (/ (+ a b c) 2.0))
(sqrt (* s (- s a) (- s b) (- s c)))
}
And now the Kermit function reads almost like the original formula.
Here Kermit behaves more like a regular programming language. In an
S-Expression, macro names need not be escaped when they are used as the
names of numeric variables.
[ [493]Top ] [ [494]Contents ] [ [495]C-Kermit Home ] [ [496]Kermit
Home ]
9.1. What is an S-Expression?
The S-Expression concept is borrowed from the Lisp programming
language. "S-Expression" is short for Symbolic Expression (itself
sometimes shortened to SEXP). S-Expressions provide a kind of
Alternative Mini-Universe within the Kermit command language when the
regular rules don't apply, a universe enclosed in parentheses.
C-Kermit does not pretend to be a full Lisp interpreter; only the
arithmetic parts of Lisp have been incorporated: S-Expressions that
operate on numbers and return numeric values (plus extensibility
features described in [497]Section 9.8, which allow some degree of
string processing).
An S-Expression is a list of zero or more items, separated by spaces,
within parentheses. Examples:
()
(1)
(a)
(+ a 1)
(* 2 a b)
If the S-Expression is empty, it has the NIL (empty) value. If it is
not empty and the first item is an operator (such as + or *), there can
be zero or more subsequent items, called the operands:
(+ 1 2)
Here the operator is "+" and the operands are "1" and "2", and the
value of the S-Expression is the value of the operation (in this case
3). The operator always comes first, which is different from the
familiar algebraic notation; this because S-Expression operators can
have different numbers of operands:
(+ 1)
(+ 1 2)
(+ 1 2 3 4 5 6 7 8 9)
If the first item in the S-Expression is not an operator, then it must
be a variable or a number (or a macro; see [498]Section 9.8), and the
S-Expression can only contain one item; in this case, the
S-Expression's value is the value of the variable or number:
(a)
(3)
Operands can be numbers, variables that have numeric values, functions
that return numbers, or other S-Expressions. To illustrate an
S-Expression within an S-Expression, observe that:
(+ 1 2)
is equivalent to any of the following (plus an infinite number of
others):
(+ 1 (+ 1 1))
(+ (- 3 2) (/ 14 (+ 3 4)))
S-Expressions can be nested to any reasonable level; for example, the
value of the following S-Expression is 64:
(- (* (+ 2 (* 3 4)) (- 9 (* 2 2))) 6)
Operators have no precedence, implied or otherwise, since they can't be
mixed. The only exceptions are unary + and -, which simply indicate the
sign of a number:
(* 3 -1)
Order of evaluation is specified entirely by parentheses, which are
required around each operator and its operands: (+ a (* b c)) instead
of (a + b * c).
S-Expressions provide a simple and isolated environment in which
Kermit's macro names can be used without the \m(...) escaping that is
normally required. Given:
define a 1
define b 2
define c 3
Then:
(+ \m(a) \m(b) \m(c))
is equivalent to:
(+ a b c)
Within an S-Expression, as in other strictly numeric contexts
([499]Section 8.2), any operand that starts with a letter is treated as
a Kermit macro name. In this context, abbreviations are not accepted;
variable names must be spelled out in full. Alphabetic case is not
significant; "a" and "A" are the same variable, but both are different
from "area".
Of course, regular Kermit variables and functions can be used in
S-Expressions in the normal ways:
(* \v(math_pi) (^ \%r 2)) ; Area of a circle with radius \%r
(+ \fjoin(&a)) ; Sum of all elements of array \&a[]
[ [500]Top ] [ [501]Contents ] [ [502]C-Kermit Home ] [ [503]Kermit
Home ]
9.2. Integer and Floating-Point-Arithmetic
Normally, if all numbers in an S-Expression are integers, the result is
an integer:
(+ 1 1) ; Result is 2
(/ 9 3) ; Result is 3
If any of the operands is floating point, however, the result is also
floating point:
(+ 1 1.0) ; Result is 2.0
(/ 9.0 3) ; Result is 3.0
If all the operands are integers but the result has a fractional part,
the result is floating point:
(/ 10 3) ; Result is 3.333333333333333
To force an integer result in such cases, use the TRUNCATE operator:
(truncate (/ 10 3)) ; Result is 3
Similarly, to force a computation to occur in floating point, you can
coerce one of its operands to FLOAT:
(+ 1 (float 1)) ; Result is 2.0
The result is also floating point if the magnitude of any integer
operand, intermediate result, or the result itself, is larger than the
maximum for the underlying machine architecture:
(^ 100 100)
If the result is too large even for floating-point representation,
"Infinity" is printed; if it is too small to be distinguished from 0,
0.0 is returned.
Large numbers can be used and large results generated, but they are
accurate only to the precision of the underlying machine. For example,
the result of:
(+ 111111111111111111111 222222222222222222222)
should be 333333333333333333333, but 333333333333333300000.0 is
produced instead if the machine is accurate to only about 16 decimal
digits, even with coercion to floating-point. The order of magnitude is
correct but the least significant digits are wrong. The imprecise
nature of the result is indicated by the ".0" at the end. Contrast
with:
(+ 111111111 222222222)
which produces an exact integer result.
[ [504]Top ] [ [505]Contents ] [ [506]C-Kermit Home ] [ [507]Kermit
Home ]
9.3. How to Use S-Expressions
S-Expressions may be given as commands to C-Kermit. Any command whose
first character is "(" (left parenthesis) is interpreted as an
S-Expression.
If you enter an S-Expression at the C-Kermit> prompt, its result is
printed:
C-Kermit>(/ 10.0 3)
3.333333333333333
C-Kermit>
If an S-Expression is executed within a macro or command file, its
value is not printed. However, you can control the printing action
with:
SET SEXPRESSION ECHO { AUTO, ON, OFF }
AUTO is the default, meaning print the value at top level only;
ON means always print the value; OFF means never print it.
In any case, the value of the most recent S-Expression (and the
S-Expression itself) may be accessed programmatically through the
following variables:
\v(sexpression)
The S-Expression most recently executed.
\v(svalue)
The value of the S-Expression most recently executed.
Besides issuing S-Expressions as commands in themselves, you can also
execute them anywhere within a Kermit command, but in this case they
must be enclosed in a function call (otherwise they are taken
literally):
\fsexpression(s)
The argument "s" is an S-Expression; the outer parentheses may
be omitted. The value of the S-Expression is returned. Note that
since S-Expressions usually contain spaces, some form of
grouping or quoting might be needed in some contexts:
echo \fsexpression((+ 1 1)) ; Outer parentheses may be included
echo \fsexpr(+ 1 1) ; Outer parentheses may be omitted
echo Value = "\fsexp(+ 1 a)" ; Can be embedded in strings
echo Value = \&a[\fsexp(/ b 2)] ; Can be used in array subscripts
if = {\fsexp(+ 1 1)} 2 { ; Braces needed here for grouping
echo One plus one still equals two
}
The IF statement illustrates how to use S-Expressions as (or in) IF or
WHILE conditions:
* Although S-Expressions and IF conditions are similar in appearance,
they are not interchangeable. Therefore you must use \fsexpr() to
let Kermit know it's an S-Expression rather than a regular IF
condition, or a boolean or algebraic expression within an IF
condition.
* In contexts where a single "word" is expected, you must enclose the
\fsexp() invocation in braces if the S-Expression contains spaces
(and most of them do).
If an S-Expression is the last command executed in a macro, its value
becomes the return value of the macro; no RETURN command is needed.
Example:
def newarea {
(let s (/ (+ \%1 \%2 \%3) 2.0))
(sqrt (* s (- s \%1) (- s \%2) (- s \%3)))
}
This is equivalent to (but more efficient than):
def newarea {
(let s (/ (+ \%1 \%2 \%3) 2.0))
return \fsexp(sqrt (* s (- s \%1) (- s \%2) (- s \%3)))
}
When an S-Expression is entered as a command -- that is, the first
nonblank character of the command is a left parenthesis -- then it is
allowed to span multiple lines, as many as you like, until the first
left parenthesis is matched:
(let s (/
(+
\%1
\%2
\%3
)
2.0
)
)
(sqrt (*
s
(- s \%1)
(- s \%2)
(- s \%3)
)
)
The S-Expression concept lends itself easily to embedding and
recursion, but the depth to which recursion can occur is limited by the
resources of the computer (memory size, address space, swap space on
disk) and other factors. There is no way that C-Kermit can know what
this limit is, since it varies not only from computer to computer, but
also from moment to moment. If resources are exhausted by recursion,
C-Kermit simply crashes; there's no way to trap this error. However,
you can set a depth limit on S-Expressions:
SET SEXPRESSION DEPTH-LIMIT number
Limits the number of times the S-Expression reader can invoke
itself without returning to the given number. The default limit
is 1000. This limit applies to S-Expressions embedded within
other S-Expressions as well as to S-Expressions that invoke
recursive macros. If the limit is exceeded, Kermit prints
"?S-Expression depth limit exceeded" and returns to its prompt.
More about recursion in [508]Section 9.8.
You can also test the depth programmatically:
\v(sdepth)
The current S-Expression invocation depth. The depth includes
both nesting level and recursion. For example, in:
(foo (foo (foo (foo (foo))))), the innermost (foo) is at depth
5.
Help, completion, and syntax checking are not available within an
S-Expression. If you type ? within an S-Expression, it says:
C-Kermit>(? S-Expression ("help sexp" for details)
As it says, typing "help sexp" will display a brief help text.
The SHOW SEXPRESSION command displays current SET SEXPRESSION settings
and related information.
[ [509]Top ] [ [510]Contents ] [ [511]C-Kermit Home ] [ [512]Kermit
Home ]
9.4. Summary of Built-in Constants and Operators
Three constants are built in:
* PI, whose value is the value of pi (the quotient of circumference
of any circle and its diameter, 3.141592653...) to the underlying
machine's precision;
* T, which always has the value 1, which signifies truth in Kermit
logical expressions or S-Expressions;
* NIL, which always has the empty value, and can serve as a False
truth value.
These constants are specific to S-Expressions and are not visible
outside them. They may not be used as the target of an assignment. So,
for example:
(setq t 0) Fails
assign t 0 Succeeds but this is not the same T!
E (the base of natural logarithms, 2.7182818184...) is not built in
since it is not intrinsic in most Lisp dialects. If you want E to be
the base of natural logarithms you can:
(setq e (exp 1))
Operators are either symbols (such as "+") or words. Words must be
spelled out in full, not abbreviated. Differences of alphabetic case
are ignored.
The most basic operation in S-Expressions is evaluation:
EVAL [ s-expression or variable or number [ another [ another ... ] ] ]
Evaluates its operands and returns the value of the last one
evaluated. Examples:
(eval) 0
(eval 1) 1
(eval a) value of a
(eval (+ 1 a)) value of a+1
(eval (setq a 1) (setq b (+ a 0.5))) value of b (= a+0.5)
You can use "." as a shorthand for EVAL:
(.)
(. 1)
(. a)
(. (+ 1 a))
(. (setq a 1) (setq b (+ a 0.5)))
Opposite of EVAL is the operator that suppresses evaluation of its
operand:
QUOTE item
The value (quote item) is "item". If the item is itself an
S-Expression, the result is the S-Expression with the outer
parentheses stripped. Examples:
(quote) (illegal)
(quote a) a
(quote hello) hello
(quote (this is a string)) this is a string
(quote this is a string) (illegal)
A shorthand notation is also accepted for quoting:
'a is equivalent to (quote a). And therefore:
'(a b c) is equivalent to (quote (a b c)).
More about quoting in [513]Section 9.8.
STRING item
Is a combination of EVAL and QUOTE. It evaluates the item as an
S-Expression, and then puts quotes around the result (more about
this in [514]Section 9.8).
The following operators assign values to variables:
SETQ [ variable [ value [ variable [ value [ ... ] ] ] ] ]
Applies to global variables. For each variable given: if a value
is not given, the variable is undefined. If a value is given,
assigns the value to the variable. The value may be a number, a
variable, or anything that resolves to a number including an
S-Expression. Returns the value of the last assignment.
Examples:
(setq) Does nothing, returns NIL.
(setq a) Undefines a, returns NIL.
(setq a 1) Assigns 1 to a, returns 1.
(setq a 1 b 2) Assigns 1 to a, 2 to b, returns 2.
(setq a 1 b 2 c) Assigns 1 to a, 2 to b, undefines c, returns NIL.
To undefine a variable that is not the final one in the list, give it a
value of "()" or NIL:
(setq a () b 2) Undefines a, assigns 2 to b, returns 2.
(setq a nil b 2) Ditto.
Note that a variable can be used right away once it has a value:
(setq a 1 b a) Assigns 1 to a, the value of a (1) to b, returns 1.
The results of SETQ (when used with macro names) can be checked
conveniently with SHOW MACRO, e.g:
show mac a b c
LET [ variable [ value [ variable [ value [ ... ] ] ] ] ]
Like SETQ, but applies to local variables. Note that "local" is
used in the Kermit sense, not the Lisp sense; it applies to the
current Kermit command level, not to the current S-Expression.
If you want to use SETQ or LET to assign a value to a backslash
variable such as \%a or \&a[2], you must double the backslash:
(setq \\%a 3)
(setq \\%b (+ \%a 1))
(setq \\&a[2] (setq (\\%c (+ \%a \%b))))
In other words:
* Double the backslash when you want to indicate the variable's NAME;
* Don't double the backslash when you want its VALUE.
See [515]Section 9.6 for a fuller explanation of variable syntax and
scope.
Here's a summary table of arithmetic operators; in the examples, a is 2
and b is -1.3:
Operator Description Example Result
+ Adds all operands (0 or more) (+ a b) 0.7
- Subtracts all operands (0 or more) (- 9 5 2 1) 1
* Multiplies all operands (0 or more) (* a (+ b 1) 3) -1.80
/ Divides all operands (2 or more) (/ b a 2) -0.325
^ Raise given number to given power (^ 3 2) 9
++ Increments variables (++ a 1.2) 3.2
-- Decrements variables (-- a) 1
ABS Absolute value of 1 operand (abs (* a b 3)) 7.8
MAX Maximum of all operands (1 or more) (max 1 2 3 4) 4
MIN Minimum of all operands (1 or more) (min 1 2 3 4) 1
MOD (%) Modulus of all operands (1 or more) (mod 7 4 2) 1
FLOAT Convert an integer to floating-point (float 1) 1.0
TRUNCATE Integer part of floating-point operand (truncate 3.333) 3
CEILING Ceiling of floating-point operand (ceiling 1.25) 2
FLOOR Floor of floating-point operand (floor 1.25) 1
ROUND Operand rounded to nearest integer (round 1.75) 2
SQRT Square root of 1 operand (sqrt 2) 1.414..
EXP e (2.71828..) to the given power (exp -1) 0.367..
SIN Sine of angle-in-radians (sin (/ pi 2)) 1.0
COS Cosine of angle-in-radians (cos pi) -1.0
TAN Tangent of angle-in-radians (tan pi) 0.0
LOG Natural log (base e) of given number (log 2.7183) 1.000..
LOG10 Log base 10 of given number (log10 1000) 3.0
The ++ and -- operators are also assignment operators and work just
like SETQ and LET in their interpretations of operators and operands,
but:
* Each target variable must already be defined and have a numeric
value;
* The assignment value is the amount by which to increment or
decrement the variable.
* If an assignment value is not given, 1 is used.
If you include more than one variable-value pair in a ++ or --
expression, every variable (except, optionally, the last) must be
followed by a value. Examples:
(++ a) Equivalent to (setq a (+ a 1)) and to (++ a 1)
(++ a 2) Equivalent to (setq a (+ a 2))
(-- a (* 2 pi)) Equivalent to (setq a (- a (* 2 pi)))
(++ a 1 b 1 c 1 d) Equivalent to four SETQs incrementing a,b,c,d by 1.
Another group of operators forms the predicates. These return a "truth
value", in which 0 (or NIL) is false, and 1 or any other nonzero number
is true.
Operator Description Example Result
= (or ==) Operands are equal (= 1 1.0) 1
!= Operands are not equal (!= 1 1.0) 0
< Operands in strictly ascending order (< 1 2 3) 1
<= Operands in ascending order (<= 1 1 2 3) 1
> Operands in strictly descending order (> 3 2 1) 1
>= Operands in descending order (<= 3 3 2 1) 1
AND (&&) Operands are all true (and 1 1 1 1 0) 0
OR (||) At least one operand is true (or 1 1 1 1 0) 1
XOR Logical Exclusive OR (xor 3 1) 0
NOT (!) Reverses truth value of operand (not 3) 0
The Exclusive OR of two values is true if one value is true and the
other value is false.
And another group operates on bits within an integer word:
Operator Description Example Result
& Bitwise AND (& 7 2) 2
| Bitwise OR (| 1 2 3 4) 7
# Bitwise Exclusive OR (# 3 1) 2
~ Reverses all bits (~ 3) -4
These operators coerce their operands to integer by truncation if
necessary. The result of bit reversal is hardware dependent.
The final category of operator works on truth values:
Operator Description Example Result
IF Conditional evaluation (if (1) 2 3) 2
IF (predicate) (s1) [ (s2) ]
The IF operator is similar to Kermit's IF command. If the
predicate is true (i.e. evaluates to a nonzero number), the
first S-Expression (s1) is evaluated and its value is returned.
Otherwise, if (s2) is given, it is evaluated and its value
returned; if (s2) is not given, nothing happens and the NIL
(empty) value is returned.
You can group multiple expressions in the s1 and s2 expressions using
EVAL (or "."):
(if (< a 0) (eval (setq x 0) (setq y 0)) (eval (setq x a) (setq y b)))
or equivalently:
(if (< a 0) (. (setq x 0) (setq y 0)) (. (setq x a) (setq y b)))
Each operator has its own requirement as to number and type of
operands. In the following table, "number" means any kind of number --
integer or floating-point -- or a variable, function, macro, or
S-Expression that returns a number; "vname" means variable name,
"fpnumber" means a floating-point number (or anything that resolves to
one), and "integer" means integer (or anything that resolves to one).
"truthvalue" means anything that resolves to a value of zero or an
empty value (which indicates false) or a nonzero value (which indicates
true). "any" means any kind of value, including none at all.
Operator Number of operands Type of operands Returns
EVAL (.) 0 or more S-Expression Last value (default NIL)
STRING 1 S-Expression string
QUOTE (') 1 word string
SETQ 0 or more vname value pairs Last value (default NIL)
LET 0 or more vname value pairs Last value (default NIL)
+ 0 or more number number (default 0)
- 0 or more number number (default 0)
* 0 or more number number (see note (1))
/ 2 or more number number
^ 2 or more number number
++ 1 or more vname value pairs Result of last increment
-- 1 or more vname value pairs Result of last decrement
ABS 1 number number
MAX 1 or more number number
MIN 1 or more number number
MOD (%) 2 number number
FLOAT 1 number fpnumber
TRUNCATE 1 number integer
CEILING 1 number integer
FLOOR 1 number integer
ROUND 1 number integer
SQRT 1 number fpnumber
EXP 1 number fpnumber
SIN 1 number fpnumber
COS 1 number fpnumber
TAN 1 number fpnumber
LOG 1 number fpnumber
LOG10 1 number fpnumber
= (==) 1 or more number truthvalue
!= 1 or more number truthvalue
< 1 or more number truthvalue
<= 1 or more number truthvalue
> 1 or more number truthvalue
>= 1 or more number truthvalue
AND (&&) 1 or more truthvalue truthvalue
OR (||) 1 or more truthvalue truthvalue
XOR 2 truthvalue truthvalue
NOT (!) 1 truthvalue truthvalue
& 1 or more number (see note 2) integer
| 1 or more number (see note 2) integer
# 2 number (see note 2) integer
~ 1 number (see note 2) integer
IF 2 or 3 truthvalue,any,any any
Operators that don't require any arguments return the default values
shown.
1. The value of "*", when used as an operand, is initially "1" and the
value of the most recent S-Expression thereafter, as in Franz Lisp.
This is handy when doing a series of calculations by hand:
C-Kermit>(* 13272.42 0.40)
5308.968
C-Kermit>(/ * 2)
2654.4840
C-Kermit>
2. The bitwise operators coerce their operands to integer by
truncation.
[ [516]Top ] [ [517]Contents ] [ [518]C-Kermit Home ] [ [519]Kermit
Home ]
9.5. Variables
As noted elsewhere in this discussion, all backslash items (variables
such as \%a, macro parameters such as \%1, array elements such as
\&a[\%i], built-in variables such as \v(ndate), built-in functions such
as \fjoin(), macro names enclosed in \m(), \s(), or \:(), etc) are
evaluated at "top level" before the S-Expression is sent to the
S-Expression reader. To use a backslash variable as the target of an
assignment (e.g. by SETQ, LET, ++, or --), you must double the
backslash, e.g. (setq \\%r 1234). This is discussed at greater length
in the next section.
Thus S-Expression reader generally deals only with macro names (not
backslash items) as variables. It is important to understand how the
reader handles macro names. There are fundamentally two kinds of
S-Expressions: those that contain a single element, such as:
(foo)
and those that contain more than one element:
(foo a b c)
If an S-Expression contains only one element, and it is the name of a
macro, the macro's definition is examined. If the definition is a
number (integer or floating-point, positive or negative), then this
becomes the value of the expression. If the definition starts with '
(apostrophe), then the quoted word or string is the value of the
expression (explained in [520]Section 9.8). Otherwise, the macro is
assumed to be composed of Kermit commands (possibly including
S-Expressions), which are executed. If the macro has a RETURN value, or
it executes an S-Expression as its last command, the result becomes the
value of the S-Expression; otherwise the result is empty.
For S-Expressions that contain more than one element, and the first
element is the name of a macro, then this macro is executed with the
arguments that are given, after the arguments are evaluated by the
S-Expression reader. Likewise, If the first element is a built-in
operator, then it is applied to the operands after they are evaluated.
In both cases, each operand is fed to the S-Expression reader
recursively for evaluation. If an operand is a number or a quoted
string, it is used as-is. But if it's a macro name, this degenerates
into the first case, and the previous paragraph applies.
Examples:
define foo 123
(foo) Result: 123
define foo 'abc
(foo) Result: abc
define foo '(one two three)
(foo) Result: one two three
define foo return \frandom(1000)
(foo) Result: 713 (or other number)
define foo (+ a b)
(foo) Result: The sum of a and b
A more difficult example:
define foo abc
(foo) Result: ???
The result in the last example depends on the definition of abc:
* If it has no definition, an error occurs; otherwise:
* If the definition is an S-Expression, the result is the
S-Expression's value; otherwise:
* If the definition consists of Kermit commands, they are executed.
But in this case "(foo)" produces the empty result, because it
doesn't RETURN anything.
The use of macros as S-Expression operators is described in
[521]Section 9.8.
[ [522]Top ] [ [523]Contents ] [ [524]C-Kermit Home ] [ [525]Kermit
Home ]
9.6. Assignments and Scope
The assignment operators SETQ and LET apply to global and local
variables, respectively. SETQ and LET are standard Lisp operators
adapted to Kermit scoping rules. When the operands are numeric or
arithmetic, SETQ is equivalent to Kermit's EVALUATE command:
(setq a (+ 1 2))
evaluate a 1 + 2
When the operand is a string, SETQ is equivalent to DEFINE:
(setq a '(this is a string))
define a this is a string
In the first case, both statements create a macro named "a" with a
value of 3. But in neither case is the macro "a" necessarily global. If
either of these commands executes in an environment (i.e. macro
invocation level) where a "local a" command has been given, the "a"
macro is global to that environment, but is not visible outside it.
LET is equivalent to the Kermit LOCAL command, followed by the
corresponding EVALUATE:
(let a (+ 1 2))
is equivalent to:
local a
evaluate a 1 + 2
Again, "local" in this context applies to the Kermit macro invocation
stack, not to the S-Expression nesting level. To illustrate, recall our
"newarea" macro:
def newarea {
(let a \%1 b \%2 c \%3)
(let s (/ (+ a b c) 2.0))
(sqrt (* s (- s a) (- s b) (- s c)))
}
Because SETQ and LET expressions return a value, they can be placed
within a larger S-Expression. In this case we can replace the first
reference to the "s" variable by its defining expression:
def newarea {
(let a \%1 b \%2 c \%3)
(sqrt (* (let s (/ (+ a b c) 2.0)) (- s a) (- s b) (- s c)))
}
This would not work if LET were local to the S-Expression, but it works
nicely in the context of Kermit macros. The previous definition is
equivalent to:
def newarea {
local a b c s
(setq a \%1 b \%2 c \%3)
(sqrt (* (setq s (/ (+ a b c) 2.0)) (- s a) (- s b) (- s c)))
}
In both cases, the variables a, b, c, and s are local to the "newarea"
macro, and global within it.
Multiple assignments can be handled in several ways. Here is the
obvious way to initialize a series of variables to the same value:
(setq a 0)
(setq b 0)
(setq c 0)
(setq s 0)
Here is a more compact and efficient way of doing the same thing:
(setq a 0 b 0 c 0 s 0)
However, in case the value was more complex, it's better to put only
one copy of it in the S-Expression; in this case we rely on the fact
that SETQ returns the value of its last assignment:
(setq a (setq b (setq c (setq s (* x (^ y 2))))))
Similarly, to set a series of variables to x, x+1, x+2, ...
(setq c (+ (setq b (+ (setq a (+ (setq s x) 1)) 1)) 1))
In the last example, you can see why "last" does not always correspond
to "rightmost" (the leftmost variable "c" is assigned last).
If you are working with backslash variables like \%a or array elements
like \&a[1], remember two rules:
1. Don't put spaces inside array brackets.
2. You must double the backslash when using SETQ, LET, ++, or -- to
assign a value to a backslash variable.
Examples of assigning to a backslash variable:
(setq x 1)
(setq \\%a 0)
(setq \\&a[x+1] 1)
(++ \\%x)
(-- \\&a[x+2])
Examples of referring to a backslash variable's value:
(setq a (+ \%a 1))
(setq b (+ \%a \&a[1]))
(++ a \%x)
(-- b \&a[1])
The special notation is required because all backslashed items (\%x
variables, array elements, built-in \v(xxx) variables, and \fxxx()
function invocations) are evaluated in a single pass BEFORE the
S-Expression is executed; any other approach would result in
unacceptable performance. So, for example, in:
declare \&a[] = 1 2 3
define \%x 4
define \%y 0
(setq \\%y (+ \%x \&a[1]))
the S-Expression becomes:
(setq \%y (+ 4 1))
before it is sent to the S-Expression evaluator. If the backslash had
not been doubled on the assignment target, the result would have been:
(setq 0 (+ 4 1))
which is illegal because you can't assign a value to a number.
Conversely, if backslashes were doubled on right-hand-side values:
(setq \\%y (+ \\%x \\&a[1])
this too, would give an error (not numeric - "\%x").
If you omit the double backslash in the assignment target, the result
depends on whether the variable already has a value:
(setq \%a (* 3 3))
If \%a has a non-numeric single-word value, then this becomes the name
of the variable that is assigned by SETQ. To illustrate:
define \%a foo
echo \%a
foo
(setq \%a (* 3 3))
echo \%a
foo
show macro foo
foo = 9
If \%a has no value, a numeric value, or a multiword value, an "invalid
assignment" error occurs.
[ [526]Top ] [ [527]Contents ] [ [528]C-Kermit Home ] [ [529]Kermit
Home ]
9.7. Conditional Expressions
The IF operator provides a compact form of decision-making within
S-Expressions. An IF expression can stand wherever a number might
stand, as long is it returns a number. Here's a quick way to obtain the
average value of all the elements in an array that contains only
numbers:
(/ (+ \fjoin(&a)) (float \fdim(&a)))
This results in a "Divide by zero" error if the array is empty. If you
want to define the average value of an empty array to be 0 instead of
getting an error, you can use IF to check the array size:
(if \fdim(&a) (/ (+ \fjoin(&a)) (float \fdim(&a))) 0)
or equivalently:
(if (not \fdim(&a)) 0 (/ (+ \fjoin(&a)) (float \fdim(&a))))
Of course, IF can fit anywhere else into an S-Expression:
(setq a (+ b (if (< c 0) 0 c)))
and the IF expression can be as complex as you like:
(setq a (+ b (if (and (or (> x 0) (> y 0)) (< c 0) (> d 1) (!= e 0)) 1 0)))
and the "then" and "else" parts can contain multiple S-Expressions
enclosed within (EVAL ...):
(if x (eval (...) (...) (...)) (eval (...) (...) (...)))
AND and OR operators are guaranteed to "short circuit". If any operand
of AND is false, none of the subsequent operands is evaluated;
likewise, if an OR operand is true, no further operands are evaluated.
Bear in mind that the S-Expression IF is not the same as Kermit IF; the
condition is only allowed to be an S-Expression or a variable or
number, not the whole list of possibilities you see when you type "if
?" at the C-Kermit> prompt. But keep reading...
[ [530]Top ] [ [531]Contents ] [ [532]C-Kermit Home ] [ [533]Kermit
Home ]
9.8. Extensibility
To extend the capabilities of S-Expressions, you can use Kermit macro
names as operators, with the following limitations:
* The macro must not have the same name as a built-in operator.
* You must use the full macro name, not an abbreviation.
And with the following enhancement:
* If the last statement executed by the macro is an S-Expression, its
value is returned automatically. In other words:
define bump (++ \%1)
is equivalent to:
define bump return \fsexpression(++ \%1)
Here's an example in which we define a FIBONACCI operator that returns
the nth element, n >= 0, of the Fibonacci series, 0 1 1 2 3 5 8 13 21
34 55, . . ., in which the first element is 0, the second is 1, and
each subsequent element is the sum of the two before it. This series
was devised by Leonardo Pisano, Filius Bonacci (Fibonacci for short) in
1202 to describe how fast rabbits can breed, and also forms the basis
for the Golden Mean, the branching behavior of plants, the spiral of a
nautilus shell, etc. (Thanks to [534]Dat Thuc Nguyen for December 2003
corrections to this section!)
We can write a FIBONACCI function as a macro easily with S-Expressions:
define FIBONACCI {
(if (== \%1 0) 0
(if (== \%1 1) 1 (+ (fibonacci (- \%1 2)) (fibonacci (- \%1 1)))))
}
You can read this as:
If the argument (\%1) is 0, return a result of 0; if it is 1,
return 1; otherwise:
return the sum of fibonacci(argument - 2) and fibonacci(argument -
1)
Note that a RETURN statement is not needed, since S-Expressions
automatically set the return value of their containing macros.
For comparison, here's how it would be coded without S-Expressions:
define FIBONACCI {
if == \%1 0 {
return 0
} else if == \%1 1 {
return 1
} else {
return \feval(\fexec(fibonacci \feval(\%1-2)) -
+ \fexec(fibonacci \feval(\%1-1)))
}
}
Now we can use the FIBONACCI function (whichever way you write it) just
as if it were a built-in operator:
(fibonacci 6)
Or:
(setq a 10)
(fibonacci a)
Within S-Expressions only (not outside them), S-Expressions themselves
can be used as macro arguments:
(setq a 2 b 4)
(setq x (fibonacci (* a b )))
The value of the S-Expression (in this case "8"), and not the
S-Expression itself, is sent to the macro.
Your macro is responsible for argument validation and error handling. A
robust Fibonacci macro would be more like this:
define FIBONACCI {
if < \v(argc) 2 end 1 ?\%0: Missing argument
if > \v(argc) 2 end 1 ?\%0: Too many arguments
if not integer \%1 end 1 ?\%0: Integers only
if < \%1 1 end 1 ?\%0: Argument out of range
(if (== \%1 0) 0
(if (== \%1 1) 1 (+ (fibonacci (- \%1 2)) (fibonacci (- \%1 1)))))
}
Recall that "END nonzero-number [ message ]" causes a macro invocation
to fail. When the macro is the operator in an S-Expression, this makes
the S-Expression fail too. Also note that our Fibonacci macro is just
an illustration, not a practical example. Since it is recursive (calls
itself), it won't work for large arguments because the call stack can
exceed available memory. See [535]Section 9.9.2 for a practical
alternative.
Kermit macros, when used as S-Expression operators, can do anything at
all except initiate file transfers: they can print messages on the
screen, read and write files, interact with the user, and so on. For
example, here's a macro ASKME that asks you to enter a number, makes
sure that you did, and then returns its value for use in the
S-Expression:
define ASKME {
local \%n
while true {
ask \%n { Number: }
if not def \%n continue
if not numeric \%n {
echo Not numeric - "\%n"
continue
}
break
}
return \%n
}
(setq a (* 2 (askme))) ; Get number from user, double it, assign result to a.
Here's a macro you can use to validate that a number is in a given
range:
define inrange {
if != \v(argc) 4 end 1 ?\%0: Wrong number of arguments
if ( < \%1 \%2 || > \%1 \%3 ) return 0
return 1
}
The first argument is the number to be checked, the second is the
minimum acceptable value, the third is the maximum. You can use this
(for example) in IF conditions:
define yes echo \%1 IS OK
define no echo \%1 IS NOT OK
(setq a -1 b 999)
(if (inrange a 0 100) (yes a) (no a))
(if (inrange b -1000 +1000) (yes b) (no b))
This is just an illustration, of course; there's already a built-in
operator to let you do range checking without help from macros:
(if (<= 0 a 100) (yes a) (no a))
(if (<= -1000 b +1000) (yes b) (no b))
To send string parameters to a macro, some kind of quoting is required
to tell the S-Expression parser to take a given "word" literally rather
than replacing it by its value. For this we use the Lisp QUOTE
operator:
define length return \flength(\%1)
(length (quote abcdefghijklmnopqrstuvwxyz))
26
This causes the string "abcdefghijklmnopqrstuvwxyz" to be sent
literally to the LENGTH macro. Kermit, like Lisp, also offers a
shortcut for QUOTE, that lets us quote a word by prefixing it with a
single quote (') character, also called apostrophe (ASCII 39):
(length 'abcdefghijklmnopqrstuvwxyz)
26
The two forms are equivalent.
How the macro treats its arguments is up to the macro. In the example
above, the argument is treated as a literal string. However, it can
also be treated as a variable name:
define string This is a string
define length return \flength(\m(\%1))
(length 'string)
16
Note the construct \m(\%1). This means "the value of the macro whose
name is the value of
\%1". The value of \%1 in this case is the word "string", and the value
of the macro whose name is "string" is "This is a string".
What if the macro takes multiple arguments, or a variable number of
them? Here's a simple macro that prints a phrase that includes its
arguments:
define complain echo It's too \%*!
(Recall that \%* means "all arguments".)
It can be called in the traditional way:
complain hot Result: "It's too hot!"
complain cold and wet Result: "It's too cold and wet!"
Or from an S-Expression if you quote the arguments:
(complain 'hot) Result: "It's too hot!"
(complain 'cold 'and 'wet) Result: "It's too cold and wet!"
To group multiple words into a single argument, use parentheses:
(complain (quote (cold and wet))) Result: "It's too cold and wet!"
(complain '(cold and wet)) Result: "It's too cold and wet!"
Note the difference:
(complain 'cold 'and 'wet) Three arguments
(complain '(cold and wet)) One argument
Since the COMPLAIN macro uses \%* to refer to all its arguments, no
matter how many, it doesn't care which form you use. But it makes a
difference in cases where the macro refers to its arguments
individually.
To illustrate, let's consider a macro that receives the name of a macro
and its argument list and executes it with its arguments, without
knowing how many arguments there are. The following LOOP macro is used
to execute the given macro with the given argument list the requested
number of times:
def loop { local i, for i 1 \%1 1 do \%2 \%3 }
Within the LOOP macro, the first argument (\%1) is the loop count, \%2
is the macro name, and \%3 is the argument list. When the LOOP macro is
invoked traditionally like this:
loop 3 complain hot
it prints "It's too hot!" three times. To invoke it from an
S-Expression, you must quote both the macro name as well as the
argument, since in this case the macro name itself is an argument:
(loop 3 'complain 'hot)
Now what if you need to send different or variable numbers of arguments
to the LOOP macro? The LOOP macro can handle it already, provided you
group the arguments into LOOP's third argument (\%3). In Kermit syntax,
without grouping:
loop 3 complain cold and wet
prints "It's too cold!" three times ("and wet" is lost); but with
grouping (either of the following two forms):
loop 3 complain {cold and wet}
loop 3 complain "cold and wet"
the LOOP macro prints "It's too cold and wet!" three times as desired.
To do the same thing in an S-Expression, just use the Lisp forms of
quoting instead of the Kermit forms; the following two are equivalent:
(loop 3 'complain (quote (cold and wet)))
(loop 3 'complain '(cold and wet))
Here's a similar example in which we write a macro that shows both the
name and the value of one or more other macros, whose names are given
as arguments (similar to "show macro"):
define display {
local \%i
for \%i 1 \v(argc)-1 1 {
echo \&_[\%i] = \m(\&_[\%i])
}
}
(Recall that \&_[] is the macro's argument vector array, equivalent to
\%1, \%2, ...) The DISPLAY macro can be used in S-Expressions like
this:
(setq a 1 b 2 c 3)
(display 'a 'b 'c 'd)
which prints:
a = 1
b = 2
c = 3
d =
The names must be quoted to prevent their evaluation before they are
sent to the macro. This ability to pass variables "by name" to macros,
rather than by value, lets you write macros that change the values of
argument variables. For example, here's a macro that doubles the value
of its argument variable:
define double (++ \%1 \%1)
which you can call like this:
(setq a 12)
(double 'a)
In the macro, \%1 is replace by the variable name "a"; "(++ a a)" adds
"a" to itself, and sets the value of "a" to the result.
There are no built-in operators other than QUOTE, ', and STRING for
handling strings in S-Expressions, but using just these, plus macros
that use Kermit's regular string-handling features, you can easily
extend S-Expressions to do string manipulation:
define len return \flen(\%1) Returns length of argument string
define cap return \fupper(\%1) Uppercase argument string
define rev return \freverse(\%1) Reverses argument string
define sub return \fsubstr(\%1,\%2,\%3) Returns substring of arg string
(len '(this is a string)) Result: 16
(rev '(this is a string)) Result: gnirts a si siht
(rev (cap '(this is a string))) Result: GNIRTS A SI SIHT
(sub (rev (cap '(this is a string))) 5 9) Result: TS A SI S
You can assign a string to a macro name as follows:
(setq foo '(this is a string))
(setq foo (quote (this is a string)))
The two are exactly equivalent. In both cases, the macro "foo" has the
value:
'(this is a string)
so when it is retrieved it can be identified as a string rather than a
number or commands to be executed. Thus:
(setq foo (quote (this is a string)))
show macro foo
foo = '(this is a string)
(foo)
this is a string
Note the different results for "show macro foo" and "(foo)". The former
shows the internal definition; the latter evaluates the variable, which
removes the quoting. And perhaps more important, note that if the
apostrophe and surrounding parentheses were not stored as part of the
definition, (foo) would try to execute "this is a string" as a command.
Given the assignment above, the following work as expected:
(len foo) Result: 16
(rev foo) Result: gnirts a si siht
(rev (cap foo)) Result: GNIRTS A SI SIHT
(sub (rev (cap foo)) 5 8) Result: TS A SI S
Note that, unlike built-in S-Expression operators that return numbers
or truth values, these operators return strings. If you want to assign
their return values to other variables, you can do so:
(setq bar (rev (cap foo))) Result: GNIRTS A SI SIHT
But now the S-Expression processor doesn't know the value of "bar" is
supposed to be a string, rather than a macro to execute. For this you
need one final special operator, STRING. The STRING operator takes an
S-Expression as an operand, evaluates it, and then returns its value
enclosed in '(), so you can use the value as a string is subsequent
S-Expressions. Use STRING for referencing macros that return strings:
(setq bar (string (rev (cap foo)))) Result: '(GNIRTS A SI SIHT)
STRING is like QUOTE, except that it evaluates its operand before
applying the quoting, rather than taking the operand literally.
To reference backslash variables or functions that return string
values, you must use the regular quoting mechanisms:
(setq time '(\v(time)))
(setq date '(\v(date)))
assign \%r this is a string
(setq s1 '(\%r))
That's because backslash items are evaluated BEFORE the S-Expression
parser ever sees them, and the values of \v(time) and so on are not
valid S-Expressions, so STRING won't like them.
Finally a brief word on the touchy topic of quoting. Suppose you want
to include (say) literal parentheses in a string that will later be
processed by the S-Expression reader (or \fsplit() or \fword()).
Normally, you can't do this because parentheses are meaningful in these
contexts. To defeat the normal parsing rules, you can quote the
parentheses with backslash. However, due to the many levels of string
processing involved, a surprisingly large amount of backslashes might
be required, for example:
(setq s '(a b (c d) \\\\\\\\\\\\\\\\(e f (g h) x\\\\\\\\\\\\\\\\) j k))
This is nearly impossible to explain(*). Instead, just remember two
points:
* In situations like this, it's better to use DEFINE to create the
string, rather than SETQ. The example above requires only double
backslashes when DEFINE is used:
define s '(a b (c d) \\(e f (g h) x\\) j k)
* The level of quoting depends on how many levels of evaluation the
string must pass through, which is not always obvious. However, the
number of backslashes required in any given situation is always a
power of 2. So if 1 doesn't work, try 2; if 2 doesn't work, try 4;
if 4 doesn't work, try 8, 16, 32, and so on.
Considerations like this apply in any scripting language (shell, Tcl,
Perl, Python, etc). The situation is known as "Quoting Hell".
(*) If you really want an explanation, here it is:
* Every SEXP has its backslash items evaluated in a single pass at
top level before being passed to the SEXP reader, so \%1,
\v(ftime), etc, can be evaluated up front, freeing the SEXP reader
of having to know about such things, which in turn makes it much
more efficient. Therefore one level of quoting is lost right away,
and therefore you must double each backslash that is to be used as
a quote.
* When the SEXP reader sees '\', it treats it as a quote; discards it
and keeps the next character. Thus '\\' becomes '\'. This would be
the end of it, except that:
* The SEXP reader must call itself recursively on its operands, so we
must double any quotes in the operands: 2^2 = 4.
* If the result is to be passed as an argument to a macro, the
backslashes must again be doubled, because the macro processor
evaluates the arguments before sending them to the macro: 2^3 = 8.
* If the macro itself is to see the quotes, rather than just the
result of the quoting, the quotes must be doubled again: 2^4 = 16.
Moral: To create string constants in which grouping characters must be
quoted, use DEFINE rather than SETQ.
[ [536]Top ] [ [537]Contents ] [ [538]C-Kermit Home ] [ [539]Kermit
Home ]
9.9. Examples
9.9.1. Statistics
The following program computes statistics -- means, maxima, minima,
variance, standard deviation, and correlation -- from data stored in
parallel arrays, \&x[] and \&y[], which can contain any mixture of
integer and floating-point numbers: positive, negative, or zero. Array
setup and validation are not shown. Except for the traditional FOR loop
and printing the results at the end, the entire computation is done
with S-Expressions:
; Initialize sums, maxima, minima, and number of elements
(setq xsum 0 ysum 0 xsum2 0 ysum2 0 xysum 0)
(setq xmin (setq xmax \&x[1]) ymin (setq ymax \&y[1]))
(setq n \fdim(&x))
; Loop through elements and accumulate sums, maxima, and minima
for i 1 n 1 {
(setq x \&x[i] y \&y[i]) ; Notational convenience
(setq xmax (max xmax x) ymax (max ymax y)) ; X and Y maxima
(setq xmin (min xmin x) ymin (min ymin y)) ; X and Y minima
(++ xsum x ysum y) ; X and Y sums
(++ xsum2 (^ x 2) ysum2 (^ y 2)) ; Sum of X and Y squares
(++ xysum (* x y)) ; Sum of XY products
}
; Calculate results
(setq xmean (/ xsum n) ymean (/ ysum n)) ; Mean X and Y
(setq xss (- xsum2 (/ (^ xsum 2) n))) ; Intermediate values
(setq yss (- ysum2 (/ (^ ysum 2) n)))
(setq xyss (- xysum (/ (* xsum ysum) n)))
(setq xvar (/ xss n) yvar (/ yss n)) ; X and Y variance
(setq sdx (sqrt xvar) sdy (sqrt yvar)) ; Std deviation in X and Y
(setq tmp (* xss yss))
(setq cc (if tmp (/ xyss (sqrt tmp)) 1.0)) ; Correlation coefficient
show macro xmean ymean xvar yvar sdx sdy cc ; Print the results
The final "if tmp" check accounts for the possibility that both arrays
contain all 0's. Results can also be printed with "echo CC = \m(cc)",
or any other desired way. Interestingly, if we had not needed the sum
of the squares and products, we could have obtained the sums, maxima,
and minima of the X's and Y's without a loop like this:
(setq xsum (+ \fjoin(&x)) ysum (+ \fjoin(&y)))
(setq xmax (max \fjoin(&x)) ymax (max \fjoin(&y)))
(setq xmin (min \fjoin(&x)) ymin (min \fjoin(&y)))
Any Kermit function that returns numbers or lists of numbers can be
included in an S-Expression as an operand.
9.9.2. Practical Fibonacci Series
The recursive Fibonacci example given previously is simple and elegant,
but not very useful since it causes memory occupation to grow each time
it calls itself, until eventually both physical memory and disk swap
space are filled and the program crashes. Even for small arguments,
like 17, execution time can be prohibitive:
(setq t1 \v(ftime))
(setq result (fibonacci 17))
(setq t2 (- \v(ftime) t1))
echo FIBONACCI(17) = \m(result): TIME = \ffpround(t2,3)
prints (on a certain rather slow computer):
FIBONACCI(17) = 1597: TIME = 5.861
Any recursive function can be recoded iteratively. The result is not as
pretty, but execution is far less expensive:
define FIBITER {
(if (== \%3 0) (\%2) (fibiter (+ \%1 \%2) \%1 (- \%3 1)))
}
define FIBONACCI {
(fibiter 1 0 \%1)
}
Here's the result on the same computer for the same argument of 17:
FIBONACCI(17) = 1597: TIME = 0.015
(47 times faster.) Execution time increases proportionally to the size
of the argument in the iterative case, whereas in the recursive case it
goes up geometrically, quickly reaching infinity.
[ [540]Top ] [ [541]Contents ] [ [542]C-Kermit Home ] [ [543]Kermit
Home ]
9.10. Differences from Algebraic Notation
In C-Kermit:
* Algebraic notation uses infix operators and normal rules of
operator precedence, with parentheses used to force exceptions to
the rules; many operations can be included in an expression.
S-Expressions use prefix operators with no intrinsic precedence;
each operation is enclosed in parentheses, and the arrangement of
parentheses determines precedence.
* Algebraic infix operators require two operands; S-Expression prefix
operators can accept a variable number of operands.
* You can use algebraic notation anywhere that C-Kermit accepts a
number, e.g. "echo \&a[((1+1)*2-1]", but you can use S-Expressions
only as top-level commands. You can, however, use either algebraic
or S-Expressions anywhere at all by enclosing them in \fevaluate()
or \fsexpression(), respectively.
* You can use any mixture of integer and floating-point numbers in
S-Expressions, but only integers are permitted in algebraic
expressions. Outside of S-Expressions, floating point arithmetic is
supported only by \ffp...() function calls.
* Operators and operands in S-Expressions must be separated by
spaces, e.g. "(+ a b)". Spaces are not required in algebraic
expressions: "((a+b)*c)".
* When assigning values to backslash variables (such as \%x or
\&a[2]) using SETQ or LET, you must double the backslash.
[ [544]Top ] [ [545]Contents ] [ [546]C-Kermit Home ] [ [547]Kermit
Home ]
9.11. Differences from Lisp
* Kermit has a lot of built-in operators not found in Lisp: ++, ^,
etc.
* Most dialects of real Lisp do not allow S-Expressions that don't
start with an operator, for example:
(a)
This expression can cause an error in Lisp (even if "a" has a
value), but is acceptable in Kermit, where it returns the value of
the variable "a". Similarly, (1) returns the value "1".
* In real Lisp, EVAL requires exactly one operand. In Kermit, it can
have 0, 1, 2, or more operands. It returns the value of the last
operand evaluated.
* Real Lisp SETQ and LET usually require an even number of operands.
Kermit allows an odd number, in which case the last (or only)
variable is undefined (i.e. deleted, destroyed).
* Kermit does not support ratios such as "7/8". Some Lisp dialects
accept ratios as numbers, and generate ratios when told to divide
two integers whose quotient is not a whole number; e.g. in Common
Lisp:
[13] USER(37): (/ (+ 1 2 3 4) 3)
10/3
[13] USER(38):
* The result of (/ 10 3) is 3.333.... Some Lisp dialects truncate the
result to 3 since both operands are integers, some don't; some give
the result as a ratio. C-Kermit always gives a floating point
result when there is a fractional part. If you want an integer
result, you can use TRUNCATE, FLOOR, or CEILING, e.g. (truncate (/
10 3)).
* There is currently no "bignum" support. Large numbers can be used
and large results generated, but (as noted in [548]Section 9.2)
they are accurate only to the precision of the underlying machine.
\v(math_precision) gives the machine precision as a number of
decimal digits, e.g. 16.
* Scientific notation for floating-point numbers is not supported. If
the magnitude of a number is greater than the precision of the
underlying hardware, the less-significant digits are shown but
their values are meaningless. If it the number is too small to be
represented internally, it is shown as "0.0".
* Many Lisp features are omitted: List processing (CAR, CDR, etc),
DEFUN, Lisp-specific control structures, and so on.
[ [549]Top ] [ [550]Contents ] [ [551]C-Kermit Home ] [ [552]Kermit
Home ]
10. FILE TRANSFER
New commands and switches:
SET TRANSFER REPORT { OFF, ON }
Enables or disables the (new) one-line message printed by Kermit
after a remote-mode file transfer to indicate the source and
destination file, complete with path, to let you know where the
file went.
SEND /TYPE:{TEXT,BINARY}
Sends only files of the given type (see [553]Section 4).
SEND /NOFOLLOWLINKS:
(UNIX only) Skip over symbolic links rather than following them
(default). This applies to wildcard and/or recursive SENDs; if a
single filename is given, and it happens to be a symbolic link,
the file it points to is sent.
SEND /FOLLOWLINKS:
(UNIX only) Follow (resolve) symbolic links. Watch out for
circular links, endless loops, etc.
SET SEND I-PACKETS { OFF, ON }
When sending commands to a Kermit server, this tells whether
command packets should be preceded by an I (information) packet,
which is used to synchronize parameters prior to executing the
command. Normally ON. The only reason to set this OFF is for
communicating with buggy Kermit servers that misbehave when an I
packet is sent to them. There is also a SET RECEIVE I-PACKETS
command, but presently it has no effect.
SET TRANSFER MESSAGE [ text ]
Sets an initial message to be shown in the Last Message field of
the fullscreen file-transfer display.
SET TRANSFER TRANSLATION { ON, OFF }
Inhibits or re-enables text-file transfer character-set
translation globally.
{ SEND, MSEND, GET, RECEIVE } /TRANSPARENT
Inhibits character-set translation for this transfer only.
{ GET, RECEIVE } /PIPES:{ON,OFF}
Overrides global TRANSFER PIPES setting for this transfer only;
ON allows incoming files with names like "!tar xf -" to be
opened as pipelines rather than regular files.
The following new "hot keys" are available when Kermit's file-transfer
display is visible:
D: Turn on debugging, open "debug.log" if not already open.
d: Turn off debugging but leave log open (if it was open).
T: Turn on debug-log timestamps.
t: Turn off debug-log timestamps.
Other improvements:
* SET FILE DOWNLOAD-DIRECTORY now works for external protocols (e.g.
sz/rz) too.
* Improved automatic per-file text/binary switching, described in
[554]Section 4.
* When sending a file group (e.g. "send *.*"), failure to open a file
is no longer fatal; now C-Kermit simply goes ahead to the next
file.
* Transaction log entries are now made for external protocols too.
[ [555]Top ] [ [556]Contents ] [ [557]C-Kermit Home ] [ [558]Kermit
Home ]
11. MODEMS AND DIALING
In C-Kermit 8.0, the default modem type for dialing has changed from
NONE (= DIRECT, meaning no modem) to GENERIC. This change should have
no impact on direct connections. For dialing, it means that, unless you
SET MODEM TYPE to a specific type, such as USROBOTICS or CONEXANT,
Kermit assumes:
1. The modem uses the Hayes AT command set.
2. The modem supports error correction, data compression, and hardware
flow control and is already configured to use them.
In fact, Kermit assumes the modem is completely configured, and
therefore does not send it an initialization string or any
configuration commands. Instead, it sends only the simplest and most
portable commands:
ATQ0V1 Give dial result codes.
ATDTnumber Dial the number.
(or ATD or ATDP, as appropriate).
The new defaults work for direct connections and for most modern modems
on most platforms, and they work much faster than "full-treatment"
dialing. If the new defaults don't work for you, or if you need to
perform explicit modem configurations or interactions, then set a
specific modem type and use the SET MODEM and SET DIAL commands as
documented in Using C-Kermit.
WARNING: Don't use the generic modem on hosts that do not support
RTS/CTS flow control. If Xon/Xoff is in use on the serial port,
you'll need to select a particular modem type so Kermit knows what
command to give it to enable Xon/Xoff flow control between itself
and your serial port.
The following new modem types were added in C-Kermit 8.0:
lucent: Lucent Venus chipset
pctel: PCTel V.90 chipset
conexant: Conexant (ex-Rockwell) modem family
zoom-v32bis: New name for "Zoom"
zoom-v34 Zoom V.34
zoom-v90 Zoom V.90 56K
zoom-v92: Zoom V.92 with V.44 data compression
zoltrix-v34: New name for "zoltrix"
zoltrix-hsp-v90: Synonym for PCTel
zoltrix-hcf-v90: Synonym for ITU-T-V250
smartlink-v90: Synonym for usrobotics (same chipset)
acer-v90: Synonym for Rockwell-v90
New DIAL-related variables:
\v(dm_hf): Dial modifier: Wait for Hook-Flash.
\v(dm_wb): Dial modifier: Wait for Bong.
Finally, if dialing fails, Kermit now prints a context-sensitive hint
suggesting possible reasons and remedies.
Added in C-Kermit 8.0.201: Rudimentary support for Caller ID, for use
with the ANSWER command. If the modem reports Caller ID information,
Kermit stores it in variables that you can access after the call is
answered:
\v(callid_date) The date of the call
\v(callid_time) The time of the call
\v(callid_name) The name of the caller
\v(callid_nmbr) The telephone number of the caller
\v(callid_mesg) A message
The format of these items depends on the originating and answering
phone companies and the modems and their configuration.
Not very many modems support Caller ID, and those that do (a) tend to
have it disabled by default, and (b) use different commands to enable
it. A quick survey shows of some current models shows:
- USR V.90: No
- ITU-T V.250: No
- Lucent Venus: No
- Diamond Supra: #CID=1
- Rockwell 56K: #CID=1
- PCTEL: #CID=1
- Zoltrix: +VCID=1
- Conexant: +VCID=1
To use Kermit's Caller ID feature, you have to set the modem to wait
for at least two rings before answering, and you have to give the
command to enable Caller ID; for example (after choosing a modem with
SET MODEM TYPE):
set modem command autoanswer on ATS0=2#CID=1\{13}
set modem command autoanswer on ATS0=2+VCID=1\{13}
These commands can be undone with:
set modem command autoanswer on ATS0=1#CID=0\{13}
set modem command autoanswer on ATS0=1+VCID=0\{13}
Kermit presently has no built-in knowledge of the Caller ID
capabilities or commands of the modems in its database.
Since the variables can be accessed only after the call is answered,
the only way to refuse a call is to answer it, inspect the variables,
and then hang it up if desired.
[ [559]Top ] [ [560]Contents ] [ [561]C-Kermit Home ] [ [562]Kermit
Home ]
12. TERMINAL CONNECTION
Now that 7-bit connections are no longer the norm, the default terminal
bytesize (also called "data size" or "word size") in C-Kermit 8.0 is 8
bits, rather than 7 bits as it was in C-Kermit 7.0 and earlier:
SET ESCAPE character
This command, which specifies your CONNECT-mode escape
character, allows you to specify any ASCII control character in
a variety of formats. C-Kermit 8.0.201 now also lets you specify
any 8-bit value, 128-255, as the escape character. In the SET
ESCAPE command, you can type the 8-bit character literally or
you can enter its numeric code. Here are examples that you can
enter from a terminal or console that uses the ISO Latin-1
character set:
C-Kermit> set escape Ã
C-Kermit> set escape 195
C-Kermit> show escape
Escape character: Code 195 (Ã): enabled
C-Kermit>
Both of these commands set the escape character value to 195
(decimal), which happens to be uppercase letter A with Tilde in
Latin-1. SHOW ESCAPE and SHOW TERMINAL show the value, as does
the CONNECT message.
SET TERMINAL AUTODOWNLOAD ERROR { STOP, CONTINUE }
When Kermit has a terminal connection to another computer, and a
file transfer is initiated automatically because a Kermit packet
was received in CONNECT mode (i.e. in the terminal screen), this
command tells what Kermit should do if the transfer fails. The
default is to STOP, which leaves Kermit in command mode with its
file-transfer display showing, so you can see that the transfer
failed and why. If you SET TERMINAL AUTODOWNLOAD ERROR CONTINUE,
this causes Kermit to return automatically to its terminal
screen (i.e. resume its CONNECT session) as if the transfer had
succeeded; this can be desirable if the entire session is under
control of a host-based script.
SET TERMINAL BYTESIZE { 7, 8 }
The byte size to use during CONNECT and INPUT command execution,
which can be more restrictive than the bytesize implied by the
current PARITY setting, but not less restrictive. In C-Kermit
7.0 and earlier, the terminal bytesize was 7 by default to
protect against the likelihood that parity was in use on the
connection without the user's knowledge. When the terminal
bytesize is 8 (as it is in C-Kermit 8.0 and later), the user
will see garbage in this (increasingly unlikely) situation. Note
that 8 data bits are required for most character sets other than
ASCII: Latin-1, UTF-8, and so on.
A new command has been added to produce timestamped session logs:
SET TERMINAL SESSION-LOG TIMESTAMPED-TEXT
Records the terminal session in text mode (like SET TERMINAL
SESSION-LOG TEXT) but adds a timestamp at the beginning of each
line. The timestamp format is hh:mm:ss.nnn, and indicates the
time at which the first character of the line appeared.
In most UNIX versions (those built with the select()-capable CONNECT
module -- pretty much all the ones that have or could have TELNET
included), an idle timeout feature has been added:
SET TERMINAL IDLE-TIMEOUT number
If the number is not 0, then Kermit is to take an action when
the given amount of time passes with no activity during CONNECT
mode. If the number is positive it is the maximum number of idle
seconds; if number is negative it represents milliseconds
(thousandths of seconds). If 0 is given as the number, there are
no idle timeouts. Synonym: SET TERMINAL IDLE-LIMIT.
SET TERMINAL IDLE-ACTION { RETURN, HANGUP, EXIT, OUTPUT [ string ] }
The action to be taken upon an idle timeout in CONNECT mode.
RETURN to the prompt, HANGUP the connection, EXIT from Kermit,
or OUTPUT the given string (if no string is given, a NUL (ASCII
0) character is sent).
SET TERMINAL IDLE-ACTION { TELNET-NOP, TELNET-AYT }
Actions that can be selected on Telnet connections only, that
might be useful if idle limits are enforced by the Telnet server
or in the TCP/IP protocol: TELNET-NOP sends a "NO Operation"
(do-nothing) command, which causes no response from the server;
TELNET-AYT sends an "Are You There" message to the server, which
should make the server send back a message. Neither of these
actions interferes with your remote session.
SET TERMINAL IDLE-ACTION is useful for connections to hosts or services
that automatically log you out after a certain amount of idle time,
e.g.:
set term idle-timeout 300
set term idle-action output \32
sends a space (as if you had pressed the space bar) every 300 seconds
(five minutes) while there is no activity (32 is the ASCII code for
space).
When C-Kermit returns from CONNECT to command mode, the reason for the
transition is given in a new variable, \v(cx_status):
0 No CONNECT command given yet.
1 User escaped back manually.
2 A trigger string was encountered.
3 IKSD entered server mode.
4 Application Program Command received from host.
5 Idle timeout.
6 Telnet protocol error.
7 Keystroke macro.
8 Time limit exceeded.
100 Internal error.
101 Carrier required by not detected.
102 I/O error on connection.
103 Disconnected by host.
104 Disconnected by user.
105 Session limit exceeded.
106 Rejected due to Telnet policy.
107 Received kill signal.
Values 100 and above indicate there is no connection.
[ [563]Top ] [ [564]Contents ] [ [565]C-Kermit Home ] [ [566]Kermit
Home ]
13. CHARACTER SETS
See the section on [567]file scanning above, and the section on
character-set conversion in [568]FTP. Also:
* True support for CP1252 (rather than treating it as Latin-1).
* Proper handling of C1 values when converting ISO 8-bit text to
UTF-8.
* TYPE /CHARACTER-SET: /TRANSLATE-TO: allows specific translations.
* The TRANSLATE command now works on multiple files.
* K_CHARSET environment variable to set the file character-set.
* SET TRANSFER TRANSLATION OFF.
* FTP client character-set translation ([569]Section 3.7).
[ [570]Top ] [ [571]Contents ] [ [572]C-Kermit Home ] [ [573]Kermit
Home ]
14. DIALOUT FROM TELNET TERMINAL SERVERS
For years, C-Kermit has supported dialing out from Telnet modem servers
(also called reverse terminal servers or access servers), but until now
there was no way for Kermit to control the communication parameters
(speed, parity, etc) on the serial port of the terminal server; it had
to use whatever was there.
But now, if you make a connection to a server that supports the Telnet
Com Port Control Option, [574]RFC 2217, you have the same degree of
control as you would have over a serial port on the computer where
Kermit is running: SET SPEED, SET FLOW, SET PARITY, SET STOP-BITS, SHOW
COMM, WAIT, SET CARRIER-WATCH, the modem-signal variables, sending
Break, and so on, apply to the connection between the terminal server
and the modem.
For example, using a Cisco Access Server 2509, where specifying a TCP
port in the 6000's selects a serial port that can be used for dialing
out:
set host xxx 6001 ; xxx is the IP hostname or address of the server
(log in if necessary) ; With a script or by hand
set modem type usr ; Tell Kermit what kind of modem it has
set speed 57600 ; This affects the server's port
set flow rts/cts ; Ditto
dial 7654321
The modem server might or might not require a login sequence. It might
also allow for automatic authentication, e.g. via Kerberos tickets.
NOTE: If the modem server requires a login sequence, then REDIAL might
not work as expected.
When you have a Telnet Com Port connection, your SET SPEED and SET FLOW
options change automatically to reflect the capabilities of the server,
rather than those of your local computer.
See the configuration manual for your server for additional
information. For example, how to set up the server to drop the Telnet
connection automatically when the telephone call is hung up (e.g.
"autohangup" on Cisco models).
For a Linux-based Telnet Com-Port server, click the Sredird link:
[ [575]Top ] [ [576]Contents ] [ [577]Sredird ] [ [578]C-Kermit Home ]
[ [579]Kermit Home ]
15. COPING WITH BROKEN KERMIT PARTNERS
There are lots of faulty Kermit protocol implementations out there,
found mainly in 3rd-party products ranging from communications software
packages to file-transfer functions embedded within devices. This topic
is covered [580]HERE for C-Kermit 7.0, but C-Kermit 8.0 adds some
additional tricks.
SET ATTRIBUTE RECORD-FORMAT { ON, OFF }
Allows control of the Kermit's Record-Format attribute. Set this
to OFF in case incoming file are refused due to unknown or
invalid record formats if you want to accept the file anyway.
SET ATTRIBUTES OFF
This is not a new trick, but it was recently discovered that the
Kermit implementation embedded within a certain kind of
punching/bending machine (Salvagnini if you must know) hangs
upon reception of standard format Kermit attributes when
receiving files. When sending files, it sends attributes of its
own, one per A-packet, which is slightly unusual but legal. When
receiving files from C-Kermit, K95, MS-DOS Kermit, Kermit-370,
etc, it simply exits upon reception of the first A-packet;
apparently it was not coded according to the protocol
specification, which allows multiple attributes per A-packet.
Solution: tell the file sender to SET ATTRIBUTES OFF.
SET SEND I-PACKETS { ON, OFF }
A Kermit server is supposed to accept I-packets; this is how the
client lets the server know its capabilities and preferences
before sending a command. Apparently there is at least one
Kermit server implementation that does not accept I-packets, and
does not properly respond with an Error packet if it gets one.
To get around such situations in C-Kermit 8.0, you can use SET
SEND I-PACKETS OFF to inhibit the sending of I packets. In this
case, the client must be able to adjust to the server's
configuration, rather than the other way around as we are used
to.
SET PROTOCOL KERMIT {} {} {}
C-Kermit 6.0 and later automatically send "autoupload" and
"autodownload" commands when in local mode and you give a file
transfer command. For example, if you tell kermit to "send
oofa.txt", Kermit sends "kermit -r" and a carriage return, in
case you had forgotten to start Kermit on the far end and told
it to receive a file. If a Kermit program had already been
started on the far end, it should harmlessly absorb this string.
However, some Kermit programs violate the Kermit protocol
definition and treat such strings as Kermit packets even though
they are not. In such cases, give this command to set the Kermit
protocol autoupload and download strings to nothing, which tells
Kermit not to send them. (This is not a new feature, but it was
not previously included in the "Coping" section of the
documentation.)
[ [581]Top ] [ [582]Contents ] [ [583]C-Kermit Home ] [ [584]Kermit
Home ]
16. NEW COMMAND-LINE OPTIONS
kermit -h Now prints a complete listing of its command-line options,
rather than an abbreviated list squeezed into a 24x80 space.
-dd Debug, like -d but adds timestamps
--version Shows C-Kermit version number.
--noperms Equivalent to SET ATTRIBUTE PROTECTION OFF.
Kermit now accepts a selection of URLs (Universal Resource Locators) as
its first command-line argument. These are:
telnet:hostname
Makes a Telnet connection to the given host (IP hostname or
address).
ftp://[user[:password]@]hostname[/path...]
Makes an FTP connection to the given host (IP hostname or
address). If a username is given, Kermit tries to log you in; if
a password is given, it is used; if not, you are prompted for
one. If no username is given, an anonymous login is performed.
If a pathname is included, Kermit tries to GET the given file.
See [585]Section 3.1.3 for details.
ftps://[user[:password]@]hostname[/path...]
Makes a secure FTP connection over SSL.
telnets://[user[:password]@]hostname
Makes a secure Telnet connection over SSL.
kermit://[user[:password]@]hostname[/path...]
Makes a connection to an [586]Internet Kermit Server.
http://[user[:password]@]hostname[/path...]
Makes a connection to Web server.
https://[user[:password]@]hostname[/path...]
Makes a connection to secure Web server.
[ [587]Top ] [ [588]Contents ] [ [589]C-Kermit Home ] [ [590]Kermit
Home ]
17. LOGS
In C-Kermit 8.0, we make an effort to keep passwords out of the debug
log. This can never be 100% effective, but it's better than before,
when there were no precautions at all. Whenever Kermit knows it's
prompting for, parsing, or transmitting a password, it temporarily
turns off logging and then turns it back on afterwards. This keeps the
debug log password-free in most common cases, but there can be no
guarantees.
As noted elsewhere, the new "-dd" command-line option selects a
timestamped debug log (equivalent to "set debug timestamps on", "log
debug debug.log").
C-Kermit 8.0 also supports a new timestamped session log via "set
session-log timestamped-text", "log session".
There have been requests for other kinds of logs, for example a command
log. These might be added at some point. One person wanted to be able
to log commands with timestamps, but only commands issued at the
prompt, not commands from files or macros, and also wanted a header
line at the beginning showing the date, user, and host. This can be
done as follows:
.filename := \v(home)commands.log ; (for example)
fopen /write \%c \m(filename)
if success {
fwrite /line \%c \v(date): User=\v(user) Host=\v(host)
fclose \%c
set debug timestamps on
log debug {| grep "CMD(P)" >> \m(filename)} append
}
[ [591]Top ] [ [592]Contents ] [ [593]C-Kermit Home ] [ [594]Kermit
Home ]
__________________________________________________________________
C-Kermit 8.0 Update Notes / [595]The Kermit Project / Columbia
University / 15 Dec 2003 - 13 Sep 2010
References
1. http://www.columbia.edu/
2. mailto:kermit@columbia.edu
3. http://www.columbia.edu/kermit/index.html
4. http://www.columbia.edu/kermit/k95.html
5. http://www.columbia.edu/kermit/ckermit.html
6. http://www.columbia.edu/kermit/ckscripts.html
7. http://www.columbia.edu/kermit/current.html
8. http://www.columbia.edu/kermit/whatsnew.html
9. http://www.columbia.edu/kermit/faq.html
10. http://www.columbia.edu/kermit/support.html
11. http://www.columbia.edu/kermit/ckermit80.html#contents
12. http://www.columbia.edu/kermit/ckermit.html
13. http://www.columbia.edu/kermit/index.html
14. http://www.amazon.com/gp/product/1555581641?ie=UTF8&tag=aleidmoreldom-20&linkCode=as2&camp=1789&creative=9325&creativeASIN=1555581641
15. http://www.columbia.edu/kermit/ckermit80.html
16. mailto:kermit-support@columbia.edu
17. http://www.columbia.edu/kermit/
18. http://www.kermit-project.org/
19. http://www.columbia.nyc.ny.us/kermit/
20. ftp://kermit.columbia.edu/kermit/f/COPYING.TXT
21. ftp://kermit.columbia.edu/kermit/f/ckcmai.c
22. http://www.columbia.edu/kermit/ckermit80.html#xv
23. http://www.columbia.edu/kermit/ck60manual.html
24. http://www.columbia.edu/kermit/ckermi70.html
25. ftp://kermit.columbia.edu/kermit/f/ckermit70.txt
26. http://www.columbia.edu/kermit/ckututor.html
27. ftp://kermit.columbia.edu/kermit/f/ckuker.nr
28. http://www.columbia.edu/kermit/security.htm
29. http://www.columbia.edu/kermit/telnet.htm
30. http://www.columbia.edu/kermit/ftpscripts.html
31. http://www.columbia.edu/kermit/ckcbwr.html
32. ftp://kermit.columbia.edu/kermit/f/ckcbwr.txt
33. http://www.columbia.edu/kermit/ckubwr.html
34. ftp://kermit.columbia.edu/kermit/f/ckubwr.txt
35. http://www.columbia.edu/kermit/ckvbwr.html
36. ftp://kermit.columbia.edu/kermit/f/ckvbwr.txt
37. http://www.columbia.edu/kermit/ckuins.html
38. ftp://kermit.columbia.edu/kermit/f/ckuins.txt
39. http://www.columbia.edu/kermit/ckvins.html
40. ftp://kermit.columbia.edu/kermit/f/ckvins.txt
41. http://www.columbia.edu/kermit/ckccfg.html
42. ftp://kermit.columbia.edu/kermit/f/ckccfg.txt
43. http://www.columbia.edu/kermit/ckcplm.html
44. ftp://kermit.columbia.edu/kermit/f/ckcplm.txt
45. http://www.columbia.edu/kermit/iksd.html
46. http://www.columbia.edu/kermit/skermit.html
47. http://www.columbia.edu/kermit/ckermit80.html#top
48. http://www.columbia.edu/kermit/ckermit.html
49. http://www.columbia.edu/kermit/index.html
50. http://www.columbia.edu/kermit/ckermit80.html#x0
51. http://www.columbia.edu/kermit/ckermit80.html#x1
52. http://www.columbia.edu/kermit/ckermit80.html#x2
53. http://www.columbia.edu/kermit/ckermit80.html#x2.1
54. http://www.columbia.edu/kermit/ckermit80.html#x2.2
55. http://www.columbia.edu/kermit/ckermit80.html#x2.2.1
56. http://www.columbia.edu/kermit/ckermit80.html#x2.2.2
57. http://www.columbia.edu/kermit/ckermit80.html#x2.2.3
58. http://www.columbia.edu/kermit/ckermit80.html#x2.2.4
59. http://www.columbia.edu/kermit/ckermit80.html#x2.2.5
60. http://www.columbia.edu/kermit/ckermit80.html#x2.2.6
61. http://www.columbia.edu/kermit/ckermit80.html#x3
62. http://www.columbia.edu/kermit/ckermit80.html#x3.1
63. http://www.columbia.edu/kermit/ckermit80.html#x3.1.1
64. http://www.columbia.edu/kermit/ckermit80.html#x3.1.2
65. http://www.columbia.edu/kermit/ckermit80.html#x3.1.3
66. http://www.columbia.edu/kermit/ckermit80.html#x3.1.4
67. http://www.columbia.edu/kermit/ckermit80.html#x3.2
68. http://www.columbia.edu/kermit/ckermit80.html#x3.3
69. http://www.columbia.edu/kermit/ckermit80.html#x3.4
70. http://www.columbia.edu/kermit/ckermit80.html#x3.5
71. http://www.columbia.edu/kermit/ckermit80.html#x3.5.1
72. http://www.columbia.edu/kermit/ckermit80.html#x3.5.2
73. http://www.columbia.edu/kermit/ckermit80.html#x3.5.3
74. http://www.columbia.edu/kermit/ckermit80.html#x3.6
75. http://www.columbia.edu/kermit/ckermit80.html#x3.6.1
76. http://www.columbia.edu/kermit/ckermit80.html#x3.6.2
77. http://www.columbia.edu/kermit/ckermit80.html#x3.6.3
78. http://www.columbia.edu/kermit/ckermit80.html#x3.7
79. http://www.columbia.edu/kermit/ckermit80.html#x3.7.1
80. http://www.columbia.edu/kermit/ckermit80.html#x3.7.2
81. http://www.columbia.edu/kermit/ckermit80.html#x3.8
82. http://www.columbia.edu/kermit/ckermit80.html#x3.9
83. http://www.columbia.edu/kermit/ckermit80.html#x3.10
84. http://www.columbia.edu/kermit/ckermit80.html#x3.10.1
85. http://www.columbia.edu/kermit/ckermit80.html#x3.10.2
86. http://www.columbia.edu/kermit/ckermit80.html#x3.10.3
87. http://www.columbia.edu/kermit/ckermit80.html#x3.11
88. http://www.columbia.edu/kermit/ckermit80.html#x4
89. http://www.columbia.edu/kermit/ckermit80.html#x5
90. http://www.columbia.edu/kermit/ckermit80.html#x6
91. http://www.columbia.edu/kermit/ckermit80.html#x6.1
92. http://www.columbia.edu/kermit/ckermit80.html#x6.2
93. http://www.columbia.edu/kermit/ckermit80.html#x6.3
94. http://www.columbia.edu/kermit/ckermit80.html#x6.4
95. http://www.columbia.edu/kermit/ckermit80.html#x6.5
96. http://www.columbia.edu/kermit/ckermit80.html#x6.6
97. http://www.columbia.edu/kermit/ckermit80.html#x7
98. http://www.columbia.edu/kermit/ckermit80.html#x8
99. http://www.columbia.edu/kermit/ckermit80.html#x8.1
100. http://www.columbia.edu/kermit/ckermit80.html#x8.2
101. http://www.columbia.edu/kermit/ckermit80.html#x8.3
102. http://www.columbia.edu/kermit/ckermit80.html#x8.4
103. http://www.columbia.edu/kermit/ckermit80.html#x8.5
104. http://www.columbia.edu/kermit/ckermit80.html#x8.6
105. http://www.columbia.edu/kermit/ckermit80.html#x8.7
106. http://www.columbia.edu/kermit/ckermit80.html#x8.8
107. http://www.columbia.edu/kermit/ckermit80.html#x8.9
108. http://www.columbia.edu/kermit/ckermit80.html#x8.10
109. http://www.columbia.edu/kermit/ckermit80.html#x8.11
110. http://www.columbia.edu/kermit/ckermit80.html#x8.12
111. http://www.columbia.edu/kermit/ckermit80.html#x8.13
112. http://www.columbia.edu/kermit/ckermit80.html#x8.14
113. http://www.columbia.edu/kermit/ckermit80.html#x9
114. http://www.columbia.edu/kermit/ckermit80.html#x9.1
115. http://www.columbia.edu/kermit/ckermit80.html#x9.2
116. http://www.columbia.edu/kermit/ckermit80.html#x9.3
117. http://www.columbia.edu/kermit/ckermit80.html#x9.4
118. http://www.columbia.edu/kermit/ckermit80.html#x9.5
119. http://www.columbia.edu/kermit/ckermit80.html#x9.6
120. http://www.columbia.edu/kermit/ckermit80.html#x9.7
121. http://www.columbia.edu/kermit/ckermit80.html#x9.8
122. http://www.columbia.edu/kermit/ckermit80.html#x9.9
123. http://www.columbia.edu/kermit/ckermit80.html#x9.10
124. http://www.columbia.edu/kermit/ckermit80.html#x9.11
125. http://www.columbia.edu/kermit/ckermit80.html#x10
126. http://www.columbia.edu/kermit/ckermit80.html#x11
127. http://www.columbia.edu/kermit/ckermit80.html#x12
128. http://www.columbia.edu/kermit/ckermit80.html#x13
129. http://www.columbia.edu/kermit/ckermit80.html#x14
130. http://www.columbia.edu/kermit/ckermit80.html#x15
131. http://www.columbia.edu/kermit/ckermit80.html#x16
132. http://www.columbia.edu/kermit/ckermit80.html#x17
133. http://www.columbia.edu/kermit/ckermit80.html#top
134. http://www.columbia.edu/kermit/ckermit.html
135. http://www.columbia.edu/kermit/index.html
136. http://www.columbia.edu/kermit/ckuins.html#x5
137. http://www.columbia.edu/kermit/ckuins.html
138. http://www.columbia.edu/kermit/ckermit80.html#x5
139. http://www.columbia.edu/kermit/ckermit80.html#x2.2
140. http://www.columbia.edu/kermit/ckermit80.html#contents
141. http://www.columbia.edu/kermit/ckermit80.html#x15
142. http://www.columbia.edu/kermit/ckermit80.html#x3.7
143. http://www.columbia.edu/kermit/ckermit80.html#ftpdates
144. http://www.columbia.edu/kermit/ckermit80.html#ftpcheck
145. http://www.columbia.edu/kermit/ckermit80.html#ftpnamelist
146. http://www.columbia.edu/kermit/ckermit80.html#srvrename
147. http://www.columbia.edu/kermit/ckermit80.html#ftpvdir
148. http://www.columbia.edu/kermit/ckermit80.html#setftptype
149. http://www.columbia.edu/kermit/ckermit80.html#x3.6
150. http://www.columbia.edu/kermit/ckermit80.html#x15
151. http://www.columbia.edu/kermit/ckermit80.html#x8.7
152. http://www.columbia.edu/kermit/ckermit80.html#x2.1
153. http://www.columbia.edu/kermit/ckermit80.html#x2.2
154. http://www.columbia.edu/kermit/ckermit80.html#x8.14
155. http://www.columbia.edu/kermit/ckermit80.html#x8.13
156. http://www.columbia.edu/kermit/ckermit80.html#x8.13
157. http://www.columbia.edu/kermit/ckututor.html
158. http://www.columbia.edu/kermit/ckuins.html
159. http://www.columbia.edu/kermit/skermit.html
160. http://www.columbia.edu/kermit/ckermit80.html#setlocus
161. http://www.columbia.edu/kermit/ckermit80.html#lcommands
162. http://www.columbia.edu/kermit/ckermit80.html#ftpuser
163. http://www.columbia.edu/kermit/ckermit80.html#showvar
164. http://www.columbia.edu/kermit/ckermit80.html#callerid
165. http://www.columbia.edu/kermit/ckermit80.html#x6.6
166. http://www.columbia.edu/kermit/ckermit80.html#x0
167. http://www.columbia.edu/kermit/ckermit80.html#x3.11
168. http://www.columbia.edu/kermit/ckermit80.html#top
169. http://www.columbia.edu/kermit/ckermit80.html#contents
170. http://www.columbia.edu/kermit/ckermit.html
171. http://www.columbia.edu/kermit/index.html
172. http://www.columbia.edu/kermit/ckermit80.html#x0
173. http://www.columbia.edu/kermit/ckermit80.html#top
174. http://www.columbia.edu/kermit/ckermit80.html#contents
175. http://www.columbia.edu/kermit/ckermit.html
176. http://www.columbia.edu/kermit/index.html
177. http://www.columbia.edu/kermit/k95.html
178. http://www.columbia.edu/kermit/sshclient.html
179. http://www.columbia.edu/kermit/skermit.html
180. http://www.columbia.edu/kermit/skermit.html
181. http://www.columbia.edu/kermit/sshclien.htm
182. http://www.columbia.edu/kermit/ckermit80.html#x3
183. ftp://ftp.isi.edu/in-notes/rfc1738.txt
184. http://www.columbia.edu/kermit/ckermit80.html#x2.2.2
185. http://www.columbia.edu/kermit/ckermit80.html#x2.2.1
186. ftp://ftp.isi.edu/in-notes/rfc2396.txt
187. ftp://ftp.isi.edu/in-notes/rfc2616.txt
188. http://www.columbia.edu/kermit/ckermit80.html#x2.2.3
189. ftp://ftp.isi.edu/in-notes/rfc2616.txt
190. http://www.columbia.edu/kermit/ckermit80.html#x8.13.7
191. http://www.columbia.edu/kermit/security.htm#x5.4
192. http://www.columbia.edu/kermit/security.htm#x15
193. http://www.columbia.edu/kermit/security.htm#x6.2
194. http://www.columbia.edu/kermit/security.html
195. http://www.columbia.edu/kermit/ckermit80.html#x16
196. http://www.columbia.edu/kermit/ckermit80.html#top
197. http://www.columbia.edu/kermit/ckermit80.html#contents
198. http://www.columbia.edu/kermit/ckermit.html
199. http://www.columbia.edu/kermit/index.html
200. http://www.columbia.edu/kermit/ckermit80.html#x3.1
201. http://www.columbia.edu/kermit/ckermit80.html#x3.2
202. http://www.columbia.edu/kermit/ckermit80.html#x3.3
203. http://www.columbia.edu/kermit/ckermit80.html#x3.4
204. http://www.columbia.edu/kermit/ckermit80.html#x3.5
205. http://www.columbia.edu/kermit/ckermit80.html#x3.6
206. http://www.columbia.edu/kermit/ckermit80.html#x3.7
207. http://www.columbia.edu/kermit/ckermit80.html#x3.8
208. http://www.columbia.edu/kermit/ckermit80.html#x3.9
209. http://www.columbia.edu/kermit/ckermit80.html#x3.10
210. http://www.columbia.edu/kermit/ckermit80.html#x3.11
211. http://www.columbia.edu/kermit/security.htm
212. http://www.columbia.edu/kermit/security.htm#servers
213. http://www.columbia.edu/kermit/ckcsets.html
214. http://www.columbia.edu/kermit/unicode.html
215. http://www.columbia.edu/kermit/ckermi70.htm#x1.5.4
216. http://www.columbia.edu/kermit/case10.html
217. http://www.columbia.edu/kermit/ckermit80.html#x4
218. http://www.columbia.edu/kermit/ckermit80.html#x3.11
219. http://www.columbia.edu/kermit/ftpscripts.html
220. http://www.columbia.edu/kermit/ckermit80.html#top
221. http://www.columbia.edu/kermit/ckermit80.html#ftp
222. http://www.columbia.edu/kermit/ftpclient.html
223. http://www.columbia.edu/kermit/ftpscripts.html
224. http://www.columbia.edu/kermit/ckermit.html
225. http://www.columbia.edu/kermit/index.html
226. http://www.columbia.edu/kermit/ckermit80.html#x3.1.1
227. http://www.columbia.edu/kermit/ckermit80.html#x3.1.3
228. http://www.columbia.edu/kermit/ckermit80.html#x3.1.4
229. http://www.columbia.edu/kermit/ckermit80.html#x3.1.3
230. http://www.columbia.edu/kermit/ckermit80.html#x3.1.3
231. http://www.columbia.edu/kermit/ckermit80.html#x3.2
232. http://www.columbia.edu/kermit/ckermit80.html#x3.5
233. http://www.columbia.edu/kermit/ckermit80.html#x3.6
234. http://www.columbia.edu/kermit/ftpscripts.html
235. http://www.columbia.edu/kermit/ckb2.htm
236. http://www.columbia.edu/kermit/ckermit80.html#ftpautolog
237. http://www.columbia.edu/kermit/ckermit80.html#ftpuser
238. http://www.columbia.edu/kermit/ckermit80.html#x3.8
239. http://www.columbia.edu/kermit/ckermit80.html#x3.8
240. http://www.columbia.edu/kermit/ckermit80.html#top
241. http://www.columbia.edu/kermit/ckermit80.html#ftp
242. http://www.columbia.edu/kermit/ckermit.html
243. http://www.columbia.edu/kermit/index.html
244. http://www.columbia.edu/kermit/ibm_ie.html
245. http://www.columbia.edu/kermit/ckermit80.html#x3.10
246. http://www.columbia.edu/kermit/ckermit80.html#top
247. http://www.columbia.edu/kermit/ckermit80.html#ftp
248. http://www.columbia.edu/kermit/ckermit.html
249. http://www.columbia.edu/kermit/index.html
250. http://www.columbia.edu/kermit/ck60manual.html
251. http://www.columbia.edu/kermit/ckermit70.html#x4.17
252. http://www.columbia.edu/kermit/ckermit70.html
253. http://www.columbia.edu/kermit/ckermit80.html#x3.6
254. http://www.columbia.edu/kermit/ckermit80.html#x3.11
255. http://www.columbia.edu/kermit/ckermit80.html#x3.1.4
256. http://www.columbia.edu/kermit/security.html
257. http://www.columbia.edu/kermit/ckermit80.html#x3.7
258. http://www.columbia.edu/kermit/ckermit80.html#x3.7
259. http://www.columbia.edu/kermit/ckermit80.html#x8.13.4
260. http://www.columbia.edu/kermit/ckermit80.html#permswitch
261. http://www.columbia.edu/kermit/ckermit80.html#ftpchmod
262. http://www.columbia.edu/kermit/ckermit80.html#x3.6.2
263. http://www.columbia.edu/kermit/ckermit80.html#x4
264. http://www.columbia.edu/kermit/ckermit80.html#top
265. http://www.columbia.edu/kermit/ckermit80.html#ftp
266. http://www.columbia.edu/kermit/ckermit.html
267. http://www.columbia.edu/kermit/index.html
268. http://www.columbia.edu/kermit/ckermit80.html#x7
269. http://www.columbia.edu/kermit/ckermit80.html#x3.8
270. http://www.columbia.edu/kermit/ckermit80.html#x3.8
271. http://www.columbia.edu/kermit/ckb2.htm
272. http://www.columbia.edu/kermit/ckermit80.html#x3.10
273. http://www.columbia.edu/kermit/ckermit80.html#x3.10
274. http://www.columbia.edu/kermit/ckermit80.html#x3.6
275. http://www.columbia.edu/kermit/ckermit80.html#setftptype
276. http://www.columbia.edu/kermit/ckermit80.html#top
277. http://www.columbia.edu/kermit/ckermit80.html#ftp
278. http://www.columbia.edu/kermit/ckermit.html
279. http://www.columbia.edu/kermit/index.html
280. http://www.columbia.edu/kermit/ckermit70.html#x4.9
281. http://www.columbia.edu/kermit/ckermit80.html#x3.5.1
282. http://www.columbia.edu/kermit/ckermit80.html#erroraction
283. http://www.columbia.edu/kermit/ckermit70.html#x1.5
284. http://www.columbia.edu/kermit/ckermit70.html#x4.7
285. http://www.columbia.edu/kermit/ckermit70.html#x1.6
286. http://www.columbia.edu/kermit/ckermit80.html#x8.13
287. http://www.columbia.edu/kermit/ckermi70.htm#x1.5.4
288. http://www.columbia.edu/kermit/ckermi70.htm
289. http://www.columbia.edu/kermit/ckermit80.html#x4
290. http://www.columbia.edu/kermit/ckermit80.html#x3.7
291. http://www.columbia.edu/kermit/ckermit80.html#x3.5.2
292. http://www.columbia.edu/kermit/ckermit80.html#x3.7
293. http://www.columbia.edu/kermit/ckermit80.html#erroraction
294. http://www.columbia.edu/kermit/ckermit80.html#x3.5.2
295. http://www.columbia.edu/kermit/ckermit80.html#erroraction
296. http://www.columbia.edu/kermit/ckermit80.html#ftpfilenames
297. http://www.columbia.edu/kermit/ckermit80.html#ftpperms
298. http://www.columbia.edu/kermit/ckermit80.html#ftpunique
299. http://www.columbia.edu/kermit/ckermit80.html#ftpfilenames
300. http://www.columbia.edu/kermit/ckermit80.html#note_utc
301. http://www.columbia.edu/kermit/ckermit80.html#note_date
302. http://www.columbia.edu/kermit/ckermit80.html#x3.6
303. http://www.boulder.nist.gov/timefreq/faq/faq.htm#10:
304. http://www.columbia.edu/kermit/ckermit80.html#x3.7
305. http://www.columbia.edu/kermit/ckermit80.html#top
306. http://www.columbia.edu/kermit/ckermit80.html#ftp
307. http://www.columbia.edu/kermit/ckermit.html
308. http://www.columbia.edu/kermit/index.html
309. http://www.columbia.edu/kermit/ckermit80.html#x3.11
310. http://www.columbia.edu/kermit/ckermi70.htm#x4.3
311. http://www.columbia.edu/kermit/ckermit70.html
312. http://www.columbia.edu/kermit/ckermit80.html#x5
313. http://www.columbia.edu/kermit/ckermit80.html#x3.6.3
314. http://www.columbia.edu/kermit/ckermit80.html#ftpfilenames
315. http://www.columbia.edu/kermit/ckermi70.htm#x4.1
316. http://www.columbia.edu/kermit/ckermi70.htm#x4.2.2
317. http://www.columbia.edu/kermit/ckermi70.htm#x1.5.4
318. http://www.columbia.edu/kermit/ckermit80.html#x3.6.2
319. http://www.columbia.edu/kermit/ckermit80.html#x3.11
320. http://www.columbia.edu/kermit/ckermit80.html#x3.11
321. http://www.columbia.edu/kermit/ckermit80.html#srvrename
322. http://www.columbia.edu/kermit/ckermi70.htm#x4.1
323. http://www.columbia.edu/kermit/ckermi70.htm
324. http://www.columbia.edu/kermit/ckb2.htm
325. http://www.columbia.edu/kermit/ckermit80.html#ftpfilenames
326. http://www.columbia.edu/kermit/ckermit80.html#x3.5.3
327. http://www.proftpd.net/
328. http://www.columbia.edu/kermit/ckermit80.html#top
329. http://www.columbia.edu/kermit/ckermit80.html#ftp
330. http://www.columbia.edu/kermit/ckermit.html
331. http://www.columbia.edu/kermit/index.html
332. http://www.columbia.edu/kermit/ckb2.htm
333. http://www.columbia.edu/kermit/ckcsets.html
334. http://www.columbia.edu/kermit/unicode.html
335. http://www.columbia.edu/kermit/ckcsets.html
336. http://www.columbia.edu/kermit/ckcsets.html
337. http://www.columbia.edu/kermit/ckermit80.html#x4
338. http://www.columbia.edu/kermit/utf8.html
339. http://www.columbia.edu/kermit/ckcsets.html
340. http://www.columbia.edu/kermit/ckermit80.html#x4
341. ftp://ftp.isi.edu/in-notes/rfc2640.txt
342. http://www.columbia.edu/kermit/ckermit80.html#top
343. http://www.columbia.edu/kermit/ckermit80.html#ftp
344. http://www.columbia.edu/kermit/ckermit.html
345. http://www.columbia.edu/kermit/index.html
346. http://www.columbia.edu/kermit/ckermit80.html#top
347. http://www.columbia.edu/kermit/ckermit80.html#ftp
348. http://www.columbia.edu/kermit/ckermit.html
349. http://www.columbia.edu/kermit/index.html
350. http://www.columbia.edu/kermit/ckermit80.html#top
351. http://www.columbia.edu/kermit/ckermit80.html#ftp
352. http://www.columbia.edu/kermit/ckermit.html
353. http://www.columbia.edu/kermit/index.html
354. http://www.columbia.edu/kermit/ftpscripts.html
355. http://www.columbia.edu/kermit/ckermit80.html#x3.2
356. http://www.columbia.edu/kermit/ckermit80.html#x3.2
357. ftp://ftp.isi.edu/in-notes/rfc959.txt
358. http://www.columbia.edu/kermit/ckscripts.html
359. http://www.columbia.edu/kermit/ckermit80.html#top
360. http://www.columbia.edu/kermit/ckermit80.html#ftp
361. http://www.columbia.edu/kermit/ftpscript.html
362. http://www.columbia.edu/kermit/ckermit.html
363. http://www.columbia.edu/kermit/index.html
364. http://www.columbia.edu/kermit/ckermit80.html#x3.11.1
365. http://www.columbia.edu/kermit/ckermit80.html#x3.11.2
366. http://www.columbia.edu/kermit/ckermit80.html#x3.11.3
367. http://www.columbia.edu/kermit/ckermit80.html#x3.11.4
368. http://www.columbia.edu/kermit/ckermit80.html#x3.11.5
369. http://www.columbia.edu/kermit/ckermit.html
370. http://www.columbia.edu/kermit/k95.html
371. http://www.columbia.edu/kermit/ckermit80.html#x3.11.5
372. ftp://ftp.isi.edu/in-notes/rfc959.txt
373. ftp://ftp.isi.edu/in-notes/rfc2389.txt
374. http://www.ietf.org/internet-drafts/draft-ietf-ftpext-mlst-16.txt
375. http://www.columbia.edu/kermit/ftpclient.html
376. http://www.columbia.edu/kermit/ckermit80.html#top
377. http://www.columbia.edu/kermit/ckermit80.html#ftp
378. http://www.columbia.edu/kermit/ckermit.html
379. http://www.columbia.edu/kermit/index.html
380. http://www.columbia.edu/kermit/ckermit80.html#x3
381. http://www.columbia.edu/kermit/ckermit80.html#ucs2
382. http://www.columbia.edu/kermit/ckermit80.html#top
383. http://www.columbia.edu/kermit/ckermit80.html#contents
384. http://www.columbia.edu/kermit/ckermit.html
385. http://www.columbia.edu/kermit/index.html
386. http://www.columbia.edu/kermit/ckermit80.html#top
387. http://www.columbia.edu/kermit/ckermit80.html#contents
388. http://www.columbia.edu/kermit/ckermit.html
389. http://www.columbia.edu/kermit/index.html
390. http://www.columbia.edu/kermit/ckb2.htm
391. http://www.columbia.edu/kermit/ckermit80.html#top
392. http://www.columbia.edu/kermit/ckermit80.html#contents
393. http://www.columbia.edu/kermit/ckermit.html
394. http://www.columbia.edu/kermit/index.html
395. http://www.columbia.edu/kermit/ckermit80.html#x4
396. http://www.columbia.edu/kermit/ckermit80.html#x4
397. http://www.columbia.edu/kermit/ckermit80.html#x8.12
398. http://www.columbia.edu/kermit/ckermit80.html#x8.1
399. http://www.columbia.edu/kermit/ckermit80.html#x12
400. http://www.columbia.edu/kermit/ckermit80.html#x8.12
401. http://www.columbia.edu/kermit/ckermit80.html#top
402. http://www.columbia.edu/kermit/ckermit80.html#contents
403. http://www.columbia.edu/kermit/ckermit.html
404. http://www.columbia.edu/kermit/index.html
405. http://www.columbia.edu/kermit/ckermit80.html#x8.14
406. http://www.columbia.edu/kermit/ckermit80.html#top
407. http://www.columbia.edu/kermit/ckermit80.html#contents
408. http://www.columbia.edu/kermit/ckermit.html
409. http://www.columbia.edu/kermit/index.html
410. http://www.columbia.edu/kermit/ckermit80.html#x9
411. http://www.columbia.edu/kermit/ckermit80.html#top
412. http://www.columbia.edu/kermit/ckermit80.html#contents
413. http://www.columbia.edu/kermit/ckermit.html
414. http://www.columbia.edu/kermit/index.html
415. http://www.columbia.edu/kermit/ckermit80.html#x8.6
416. http://www.columbia.edu/kermit/ckermit80.html#top
417. http://www.columbia.edu/kermit/ckermit80.html#contents
418. http://www.columbia.edu/kermit/ckermit.html
419. http://www.columbia.edu/kermit/index.html
420. http://www.columbia.edu/kermit/ckermit80.html#top
421. http://www.columbia.edu/kermit/ckermit80.html#contents
422. http://www.columbia.edu/kermit/ckermit.html
423. http://www.columbia.edu/kermit/index.html
424. http://www.columbia.edu/kermit/ckermit80.html#top
425. http://www.columbia.edu/kermit/ckermit80.html#contents
426. http://www.columbia.edu/kermit/ckermit.html
427. http://www.columbia.edu/kermit/index.html
428. http://www.columbia.edu/kermit/ckermit80.html#fjoin
429. http://www.columbia.edu/kermit/ckermit80.html#fsplit
430. http://www.columbia.edu/kermit/ckermit80.html#x8.10
431. http://www.columbia.edu/kermit/ckermit80.html#top
432. http://www.columbia.edu/kermit/ckermit80.html#contents
433. http://www.columbia.edu/kermit/ckermit.html
434. http://www.columbia.edu/kermit/index.html
435. http://www.columbia.edu/kermit/ckermit80.html#x9
436. http://www.columbia.edu/kermit/ckermit80.html#x9
437. http://www.columbia.edu/kermit/ckermit80.html#x9
438. http://www.columbia.edu/kermit/ckermit80.html#x3
439. http://www.columbia.edu/kermit/ckermit80.html#x3
440. http://www.columbia.edu/kermit/ckermit80.html#x3.2
441. http://www.columbia.edu/kermit/ckermit80.html#x3.2
442. http://www.columbia.edu/kermit/ckermit80.html#x3.8
443. http://www.columbia.edu/kermit/ckermit80.html#x3
444. http://www.columbia.edu/kermit/ckermit80.html#x3
445. http://www.columbia.edu/kermit/ckermit80.html#x3
446. http://www.columbia.edu/kermit/ckermit80.html#x3.2
447. http://www.columbia.edu/kermit/ckermit80.html#x3
448. http://www.columbia.edu/kermit/ckermit80.html#x8.13
449. http://www.columbia.edu/kermit/ckermit80.html#x8.13
450. http://www.columbia.edu/kermit/ckermit80.html#x9
451. http://www.columbia.edu/kermit/ckermit80.html#x8.10
452. http://www.columbia.edu/kermit/ckermit80.html#x8.7.4
453. http://www.columbia.edu/kermit/ckermit80.html#top
454. http://www.columbia.edu/kermit/ckermit80.html#contents
455. http://www.columbia.edu/kermit/ckermit.html
456. http://www.columbia.edu/kermit/index.html
457. http://www.columbia.edu/kermit/ckermit80.html#top
458. http://www.columbia.edu/kermit/ckermit80.html#contents
459. http://www.columbia.edu/kermit/ckermit.html
460. http://www.columbia.edu/kermit/index.html
461. http://www.columbia.edu/kermit/ckermit80.html#top
462. http://www.columbia.edu/kermit/ckermit80.html#contents
463. http://www.columbia.edu/kermit/ckermit.html
464. http://www.columbia.edu/kermit/index.html
465. http://www.columbia.edu/kermit/ckermit80.html#x8.7
466. http://www.columbia.edu/kermit/ckermit80.html#top
467. http://www.columbia.edu/kermit/ckermit80.html#contents
468. http://www.columbia.edu/kermit/ckermit.html
469. http://www.columbia.edu/kermit/index.html
470. http://www.columbia.edu/kermit/ckermit80.html#scriptedit
471. http://www.columbia.edu/kermit/ckermit80.html#top
472. http://www.columbia.edu/kermit/ckermit80.html#contents
473. http://www.columbia.edu/kermit/ckermit.html
474. http://www.columbia.edu/kermit/index.html
475. http://www.columbia.edu/kermit/ckermit80.html#top
476. http://www.columbia.edu/kermit/ckermit80.html#contents
477. http://www.columbia.edu/kermit/ckermit.html
478. http://www.columbia.edu/kermit/index.html
479. ftp://ftp.isi.edu/in-notes/rfc2822.txt
480. ftp://ftp.isi.edu/in-notes/rfc2822.txt
481. ftp://ftp.isi.edu/in-notes/rfc2822.txt
482. ftp://ftp.isi.edu/in-notes/rfc2822.txt
483. http://www.columbia.edu/kermit/ckermit80.html#top
484. http://www.columbia.edu/kermit/ckermit80.html#contents
485. http://www.columbia.edu/kermit/ckermit.html
486. http://www.columbia.edu/kermit/index.html
487. http://www.columbia.edu/kermit/ckermit80.html#x8.1
488. http://www.columbia.edu/kermit/ckermit80.html#top
489. http://www.columbia.edu/kermit/ckermit80.html#contents
490. http://www.columbia.edu/kermit/ckermit.html
491. http://www.columbia.edu/kermit/index.html
492. http://www.columbia.edu/kermit/ckermit80.html#x8.2
493. http://www.columbia.edu/kermit/ckermit80.html#top
494. http://www.columbia.edu/kermit/ckermit80.html#contents
495. http://www.columbia.edu/kermit/ckermit.html
496. http://www.columbia.edu/kermit/index.html
497. http://www.columbia.edu/kermit/ckermit80.html#x9.8
498. http://www.columbia.edu/kermit/ckermit80.html#x9.8
499. http://www.columbia.edu/kermit/ckermit80.html#x8.2
500. http://www.columbia.edu/kermit/ckermit80.html#top
501. http://www.columbia.edu/kermit/ckermit80.html#contents
502. http://www.columbia.edu/kermit/ckermit.html
503. http://www.columbia.edu/kermit/index.html
504. http://www.columbia.edu/kermit/ckermit80.html#top
505. http://www.columbia.edu/kermit/ckermit80.html#contents
506. http://www.columbia.edu/kermit/ckermit.html
507. http://www.columbia.edu/kermit/index.html
508. http://www.columbia.edu/kermit/ckermit80.html#x9.8
509. http://www.columbia.edu/kermit/ckermit80.html#top
510. http://www.columbia.edu/kermit/ckermit80.html#contents
511. http://www.columbia.edu/kermit/ckermit.html
512. http://www.columbia.edu/kermit/index.html
513. http://www.columbia.edu/kermit/ckermit80.html#x9.8
514. http://www.columbia.edu/kermit/ckermit80.html#x9.8
515. http://www.columbia.edu/kermit/ckermit80.html#x9.6
516. http://www.columbia.edu/kermit/ckermit80.html#top
517. http://www.columbia.edu/kermit/ckermit80.html#contents
518. http://www.columbia.edu/kermit/ckermit.html
519. http://www.columbia.edu/kermit/index.html
520. http://www.columbia.edu/kermit/ckermit80.html#x9.8
521. http://www.columbia.edu/kermit/ckermit80.html#x9.8
522. http://www.columbia.edu/kermit/ckermit80.html#top
523. http://www.columbia.edu/kermit/ckermit80.html#contents
524. http://www.columbia.edu/kermit/ckermit.html
525. http://www.columbia.edu/kermit/index.html
526. http://www.columbia.edu/kermit/ckermit80.html#top
527. http://www.columbia.edu/kermit/ckermit80.html#contents
528. http://www.columbia.edu/kermit/ckermit.html
529. http://www.columbia.edu/kermit/index.html
530. http://www.columbia.edu/kermit/ckermit80.html#top
531. http://www.columbia.edu/kermit/ckermit80.html#contents
532. http://www.columbia.edu/kermit/ckermit.html
533. http://www.columbia.edu/kermit/index.html
534. mailto:thucdat@hotmail.com
535. http://www.columbia.edu/kermit/ckermit80.html#x9.9.2
536. http://www.columbia.edu/kermit/ckermit80.html#top
537. http://www.columbia.edu/kermit/ckermit80.html#contents
538. http://www.columbia.edu/kermit/ckermit.html
539. http://www.columbia.edu/kermit/index.html
540. http://www.columbia.edu/kermit/ckermit80.html#top
541. http://www.columbia.edu/kermit/ckermit80.html#contents
542. http://www.columbia.edu/kermit/ckermit.html
543. http://www.columbia.edu/kermit/index.html
544. http://www.columbia.edu/kermit/ckermit80.html#top
545. http://www.columbia.edu/kermit/ckermit80.html#contents
546. http://www.columbia.edu/kermit/ckermit.html
547. http://www.columbia.edu/kermit/index.html
548. http://www.columbia.edu/kermit/ckermit80.html#x9.2
549. http://www.columbia.edu/kermit/ckermit80.html#top
550. http://www.columbia.edu/kermit/ckermit80.html#contents
551. http://www.columbia.edu/kermit/ckermit.html
552. http://www.columbia.edu/kermit/index.html
553. http://www.columbia.edu/kermit/ckermit80.html#x4
554. http://www.columbia.edu/kermit/ckermit80.html#x4
555. http://www.columbia.edu/kermit/ckermit80.html#top
556. http://www.columbia.edu/kermit/ckermit80.html#contents
557. http://www.columbia.edu/kermit/ckermit.html
558. http://www.columbia.edu/kermit/index.html
559. http://www.columbia.edu/kermit/ckermit80.html#top
560. http://www.columbia.edu/kermit/ckermit80.html#contents
561. http://www.columbia.edu/kermit/ckermit.html
562. http://www.columbia.edu/kermit/index.html
563. http://www.columbia.edu/kermit/ckermit80.html#top
564. http://www.columbia.edu/kermit/ckermit80.html#contents
565. http://www.columbia.edu/kermit/ckermit.html
566. http://www.columbia.edu/kermit/index.html
567. http://www.columbia.edu/kermit/ckermit80.html#x4
568. http://www.columbia.edu/kermit/ckermit80.html#x3.7
569. http://www.columbia.edu/kermit/ckermit80.html#x3.7
570. http://www.columbia.edu/kermit/ckermit80.html#top
571. http://www.columbia.edu/kermit/ckermit80.html#contents
572. http://www.columbia.edu/kermit/ckermit.html
573. http://www.columbia.edu/kermit/index.html
574. ftp://ftp.isi.edu/in-notes/rfc2217.txt
575. http://www.columbia.edu/kermit/ckermit80.html#top
576. http://www.columbia.edu/kermit/ckermit80.html#contents
577. ftp://kermit.columbia.edu/kermit/sredird/
578. http://www.columbia.edu/kermit/ckermit.html
579. http://www.columbia.edu/kermit/index.html
580. http://www.columbia.edu/kermit/ckermi70.htm#x4.22
581. http://www.columbia.edu/kermit/ckermit80.html#top
582. http://www.columbia.edu/kermit/ckermit80.html#contents
583. http://www.columbia.edu/kermit/ckermit.html
584. http://www.columbia.edu/kermit/index.html
585. http://www.columbia.edu/kermit/ckermit80.html#x3.1.3
586. http://www.columbia.edu/kermit/cuiksd.html
587. http://www.columbia.edu/kermit/ckermit80.html#top
588. http://www.columbia.edu/kermit/ckermit80.html#contents
589. http://www.columbia.edu/kermit/ckermit.html
590. http://www.columbia.edu/kermit/index.html
591. http://www.columbia.edu/kermit/ckermit80.html#top
592. http://www.columbia.edu/kermit/ckermit80.html#contents
593. http://www.columbia.edu/kermit/ckermit.html
594. http://www.columbia.edu/kermit/index.html
595. http://www.columbia.edu/kermit/index.html