I have a problem with the Korn shell option that lets you edit and reuse
commands: "set -o vi". Normally, if you type this in a Korn-shell, you can
edit and reuse previous commands (see "A/UX User Interface" page 3-8) by
pressing Escape and then "+" or "-" to scroll through your commands and edit
them using vi-commands.
This works perfectly well except in one case. If you go to multi-user ("init
2") and log on as "rootksh", you can't go back to previous commands (after a
"set -o vi"). You still have to type "i" (after an Escape) before you can enter
commands. This means that the keyboard input is intercepted by that pseudo-vi
editor and not all functions are allowed, although you should be in a normal
Korn-shell. If you try to do "exec ksh" from there, you will have the same
problem in that shell, but everything works fine if you start another Korn
shell ("ksh"). If you do a "su rootksh", you don't have the problem.
You only have the problem (as far as I tested it) when you go to multi-user and
log on as "rootksh". This seems to generate a shell that doesn't behave like
other Korn shells. Is this normal? Is this a known Korn shell problem (I
didn't test this on other UNIX machines)? Is this
typical for A/UX?
We have verified the problem that the particular "rootksh" login account cannot
perform the Korn shell previous command function by issuing Escape and "-".