David Morano wrote:
> I am anxious to get an officially ported (meaning: OpenSolaris) AST-KSH-93
> to the Solaris platform (with KSH command builtins eventually replacing
> the corresponding existing Solaris commands), but I am not so eager to
> see various feature enhancements to KSH be done quickly (in the space of
> less than a year's worth of thinking or so) just because we are finally
> starting to get excited about the use of KSH on Solaris. I've been
> excited about KSH since like 1983, so I would rather see progress be made
> in the area of replacing Solaris program commands with KSH builtins.
> I'd also like to see the Solaris /usr/bin/sh and /sbin/sh replaced
> with AST-KSH-93. Also, how about replacing the Solaris /usr/bin/ksh
> with AST-KSH-93 (if you get my drift)? There is even /usr/dt/bin/dtksh
> that could suffer being updated someday! There seems to be plenty of
> work to do before we try to quckly rush in some totally new features.
David,
Your exposure to ksh goes back to the first release of AT&T SV?
If you search the archives you can find the current status of several of
these ideas. Here's a summary of a few of them.
* /sbin and /lib contain the minimal set of components required to mount
/usr if it would be separate. ksh93 in there is overkill.
* /etc/rc scripts should use {/usr}/bin/sh after /usr is available
Pretty moot with SMF - but SMF could use ksh93.
* dtksh is part of CDE, which is still closed source - dead
end for now. More productive would be a gksh or Kksh :-),
or a g/K-agnostic toolkit.
Cheers,
Henk
--
Henk Langeveld <henk at hlangeveld.nl>
