This is primarily addressed to Roland, but is applicable to anyone else involved in the /sbin/sh -> ksh93 effort.
What is the plan to deal with compatibility between the existing /sbin/sh and ksh93? As an example, recently it was discovered that Solaris /usr/bin/printf allows a certain syntax that it shouldn't according to the UNIX specification. Since ksh93 has "built in" commands for many of these, POSIX behaviour is often enforced. As the evaluation of ksh93 as a replacement for /bin/sh progresses, I expect there to be a lot more cases where ksh93 (rightfully) causes some programs to break due to non-POSIX compliance in the current Solaris environment. In the case of bug 502, obviously I don't expect ksh93 to change to accommodate broken programs; especially since, according to Roland, upcoming VSW&Co.(?) changes are going to cause Solaris /usr/bin/printf to break some programs as well. However, there will be other cases where the breakage is far more severe, and isn't because of something wrong the programmer did. It will likely be something they had to do a certain way because of the way the existing /sbin/sh worked. As such, what is the general plan or approach for dealing with compatibility issues as they arise? At the very least, each compatibility problems needs to be documented and considered during ARC, etc., obviously. Cheers, -- Shawn Walker, Software and Systems Analyst http://binarycrusader.blogspot.com/ "To err is human -- and to blame it on a computer is even more so." - Robert Orben
