[ksh93-integration-discuss] Re: /etc/ksh.kshrc for ksh93[PSARC/2006/587 Timeout: 10/24/2006]

[Don Cragun]

All,
	This is a status report on this case from the case owner...

1.  login.1 and su.1 man pages have been added to the materials
    directory.  Diffmarks in login.1 show changes made since PSARC case
    2006/550 was approved.  Diffmarks in su.1 show changes from the
    current man page (su.1 was not updated by 2006/550).  These changes
    are supplied in response to a question raised by John Plocher.

2.  The interfaces table is updated as follows:
	Interface	Stability	Description
	=========       =========	===========
	/etc/ksh.kshrc	Committed 	Per-system configuration file
					for interactive ksh93 sessions

	/etc/ksh.kshrc	Uncommitted 
	contents

3.  There is a lot of discussion about whether any default editing mode
    should be provided for ksh93 and if so:
    A.	which one,
    B.  where it should be set, and
    C.  should it also be set for /usr/bin/ksh and /usr/xpg4/bin/sh.

    My personal interpretation of the discussion so far is that the
    vast majority of people participating in the discussion believe
    that there should be a default, and:
    A.  Which one isn't really an architectural issue.  Some Linux
	distros default to none, some default to gmacs, and some
	default to vi.  Most other shells on Linux distros default to
	a gmacs/emacs-like editing mode.
    B.  /etc/ksh.kshrc and $HOME/.kshrc are the only places to do this
	without also affecting /usr/bin/ksh and /usr/xpg4/bin/sh users.
	The project team has done a lot of research in this area and
	believes that /etc/ksh.kshrc is a better choice than
	$HOME/.kshrc.  Although I think either one would be reasonable,
	I see no reason for the ARC to override this decision.  I don't
	see it as architecture.  A user can easily override the default
	in a personal $HOME/.kshrc or $HOME/.profile no matter which
	way this is handled.
    C.	If (or when) a case comes forward to replace one or more of
	/sbin/sh, /usr/bin/sh, /usr/bin/ksh, and /usr/xpg4/bin/sh with
	ksh93 we will need to discuss this again as a compatibility
	issue.  That discussion should not sidetrack this case.  Note
	that /sbin/sh and /usr/bin/sh do not currently provide any
	command line editing options.
	
	I do not believe this case should change the behavior of
	/usr/bin/ksh nor /usr/xpg4/bin/sh.  Note that this case is
	seeking a patch binding with a possibility of being backported
	to prior releases.  I do not believe that a patch that installs
	an alternative shell as part of a feature patch should affect
	the behavior of other shells.  The project team expects to
	integrate the changes for PSARC case 2006/550 and this case at
	the same time.

	Sincerely,
	Don