The required option argument -t identifies the table to be used for conversion. If the table has already been loaded as a shared table (see kbdload(1M)) it is attached. If, however, the table has not been loaded, an attempt is made to load it. If the specified table name is not an absolute pathname then the name of the system mapping library is prepended to the argument, and an attempt is made to load the table from the resulting pathname (that is, it becomes an argument to the loader, kbdload). Assuming the table can be loaded, it is attached.
The argument to -f defines the file from which the table will be loaded, overriding the default action described above. The file is loaded (in its entirety), and the named table attached. This option should be used if the default action would fail.
The output file specified by -o must not already exist (a safety feature). The option -F may be used to override the check for existence of the output file; in this case, any existing outfile will be truncated before being written.
The following example loads the Dvorak table from a
different file, then converts standard input to standard output.
The Dvorak table (assumed to be non-resident) is explicitly
loaded from an absolute path beginning at the user's home directory:
kbdpipe -t Dvorak -f $HOME/tables/Dvorak.tab
Users may use kbd tables in programs at user level by opening a pipe, pushing the module, and setting via related commands. Therefore, there is no need to use the kbdpipe command. kbdpipe may not be supported in future releases.