tset, reset - terminal initialization
tset [-IQVcqrsw] [-] [-e ch] [-i ch] [-k ch] [-m mapping] [terminal] reset [-IQVcqrsw] [-] [-e ch] [-i ch] [-k ch] [-m mapping] [terminal]
Tset initializes terminals. Tset first determines the type of terminal that you are using. This determination is done as follows, using the first terminal type found.
1. The terminal argument specified on the command line.
2. The value of the TERM environmental variable.
3. (BSD systems only.) The terminal type associated with the standard error output device in the /etc/ttys file. (On Linux and System-V-like UNIXes, getty does this job by setting TERM according to the type passed to it by /etc/inittab.)
4. The default terminal type, ‘’unknown’’.
If the terminal type was not specified on the command-line, the -m option mappings are then applied (see the section TERMINAL TYPE MAPPING for more information). Then, if the terminal type begins with a question mark (’’?’’), the user is prompted for confirmation of the terminal type. An empty response confirms the type, or, another type can be entered to specify a new type. Once the terminal type has been determined, the terminfo entry for the terminal is retrieved. If no terminfo entry is found for the type, the user is prompted for another terminal type.
Once the terminfo entry is retrieved, the window size, backspace, interrupt and line kill characters (among many other things) are set and the terminal and tab initialization strings are sent to the standard error output. Finally, if the erase, interrupt and line kill characters have changed, or are not set to their default values, their values are displayed to the standard error output. Use the -c or -w option to select only the window sizing versus the other initialization. If neither option is given, both are assumed.
When invoked as reset, tset sets cooked and echo modes, turns off cbreak and raw modes, turns on newline translation and resets any unset special characters to their default values before doing the terminal initialization described above. This is useful after a program dies leaving a terminal in an abnormal state. Note, you may have to type
(the line-feed character is normally control-J) to get the terminal to work, as carriage-return may no longer work in the abnormal state. Also, the terminal will often not echo the command.
The options are as follows:
The arguments for the -e, -i, and -k options may either be entered as actual characters or by using the ‘hat’ notation, i.e. control-h may be specified as ‘’^H’’ or ‘’^h’’.
It is often desirable to enter the terminal type and information about the terminal’s capabilities into the shell’s environment. This is done using the -s option.
When the -s option is specified, the commands to enter the information into the shell’s environment are written to the standard output. If the SHELL environmental variable ends in ‘’csh’’, the commands are for csh, otherwise, they are for sh. Note, the csh commands set and unset the shell variable noglob, leaving it unset. The following line in the .login or .profile files will initialize the environment correctly:
eval ‘tset -s options ... ‘
When the terminal is not hardwired into the system (or the current system information is incorrect) the terminal type derived from the /etc/ttys file or the TERM environmental variable is often something generic like network, dialup, or unknown. When tset is used in a startup script it is often desirable to provide information about the type of terminal used on such ports.
The purpose of the -m option is to map from some set of conditions to a terminal type, that is, to tell tset ‘’If I’m on this port at a particular speed, guess that I’m on that kind of terminal’’.
The argument to the -m option consists of an optional port type, an optional operator, an optional baud rate specification, an optional colon (’’:’’) character and a terminal type. The port type is a string (delimited by either the operator or the colon character). The operator may be any combination of ‘’>’’, ‘’<’’, ‘’@’’, and ‘’!’’; ‘’>’’ means greater than, ‘’<’’ means less than, ‘’@’’ means equal to and ‘’!’’ inverts the sense of the test. The baud rate is specified as a number and is compared with the speed of the standard error output (which should be the control terminal). The terminal type is a string.
If the terminal type is not specified on the command line, the -m mappings are applied to the terminal type. If the port type and baud rate match the mapping, the terminal type specified in the mapping replaces the current type. If more than one mapping is specified, the first applicable mapping is used.
For example, consider the following mapping: dialup>9600:vt100. The port type is dialup , the operator is >, the baud rate specification is 9600, and the terminal type is vt100. The result of this mapping is to specify that if the terminal type is dialup, and the baud rate is greater than 9600 baud, a terminal type of vt100 will be used.
If no baud rate is specified, the terminal type will match any baud rate. If no port type is specified, the terminal type will match any port type. For example, -m dialup:vt100 -m :?xterm will cause any dialup port, regardless of baud rate, to match the terminal type vt100, and any non-dialup port type to match the terminal type ?xterm. Note, because of the leading question mark, the user will be queried on a default port as to whether they are actually using an xterm terminal.
No whitespace characters are permitted in the -m option argument. Also, to avoid problems with meta-characters, it is suggested that the entire -m option argument be placed within single quote characters, and that csh users insert a backslash character (’’\’’) before any exclamation marks (’’!’’).
The tset command appeared in BSD 3.0. The ncurses implementation was lightly adapted from the 4.4BSD sources for a terminfo environment by Eric S. Raymond <firstname.lastname@example.org>.
The tset utility has been provided for backward-compatibility with BSD environments (under most modern UNIXes, /etc/inittab and getty(1) can set TERM appropriately for each dial-up line; this obviates what was tset’s most important use). This implementation behaves like 4.4BSD tset, with a few exceptions specified here.
The -S option of BSD tset no longer works; it prints an error message to stderr and dies. The -s option only sets TERM, not TERMCAP. Both these changes are because the TERMCAP variable is no longer supported under terminfo-based ncurses, which makes tset -S useless (we made it die noisily rather than silently induce lossage).
There was an undocumented 4.4BSD feature that invoking tset via a link named ‘TSET’ (or via any other name beginning with an upper-case letter) set the terminal to use upper-case only. This feature has been omitted.
The -A, -E, -h, -u and -v options were deleted from the tset utility in 4.4BSD. None of them were documented in 4.3BSD and all are of limited utility at best. The -a, -d, and -p options are similarly not documented or useful, but were retained as they appear to be in widespread use. It is strongly recommended that any usage of these three options be changed to use the -m option instead. The -n option remains, but has no effect. The -adnp options are therefore omitted from the usage summary above.
It is still permissible to specify the -e, -i, and -k options without arguments, although it is strongly recommended that such usage be fixed to explicitly specify the character.
As of 4.4BSD, executing tset as reset no longer implies the -Q option. Also, the interaction between the - option and the terminal argument in some historic implementations of tset has been removed.
The tset command uses the SHELL and TERM environment variables.
system port name to terminal type mapping database (BSD versions only).
terminal capability database
csh(1) , sh(1) , stty(1) , setupterm(3) , tty(4) , termcap(5) , ttys(5) , environ(7)
Table of Contents