I. Szczesniak wrote:
>
>
> I don't understand why ksh88 would be needed in a *new* distribution.
> ksh88 is buggy and the Solaris derivate is not compatible to the
> original ksh88i from David Korn nor the ksh88 versions on AIX or
> HP/UX. All newer operating systems like Linux, SCO or AIX6 switched by
> ksh93 as /bin/ksh.
>   
Solaris has had ksh93 available for a long time, but not in an obvious 
place.  Under  /usr/dt (The old CDE area) you will find dtksh which is 
an extension of an early release of ksh93.  It also has limitations (I 
think the arrays can not exceed 4K elements)  The fact that /usr/bin/ksh 
is not based on ksh93 was due to the licensing limitation  put in place 
at the time by AT&T/Bell Labs.  It was allowed in the CDE release area 
due to agreements made by all the parties which created CDE in the first 
place.


> In my opinion it does not make sense to retain ksh88 as /bin/ksh in Indiana.
>   
I have done a lot of ksh scripting (much of it when I worked for Sun) 
and my experience has shown that ksh93 is upward compatible with ksh88.  
but the fact is there are some Solaris utility scripts were written to 
use /usr/bin/ksh and I think nobody wanted to track them all down and 
verify they all ran  with a ksh93 version. 

Perhaps for a project like Indiana, we can get away with switching to 
ksh92 as /usr/bin/ksh and live with anyone who is stuck because they 
rely on some strange corner case which ksh93 does not implement right.

Sam Nuwayser

>   
_______________________________________________
indiana-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/indiana-discuss

Reply via email to