tech-userlevel archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: mksh import

If I understand correctly, I believe someone has suggested that we replace the existing /bin/sh in the base with a more fully featured shell such as ksh93 / mksh.
I would second and support such a move. Unless someone can show that
ksh93 / mksh would adversely affect the performance of the system in a significant way then I can think of no reason why this should not be done. By providing ksh93 / mksh as /bin/sh in the base it provides a guarantee to developers that they are able to take advantage of the most powerful and efficient features that a modern shell has to offer.
In doing so NetBSD will be a better system for it.

I would recommend that David Korn's ksh93 be used as the default shell in the base. Mr Korn actively maintains and improves ksh93 and has done so for over 20 years. The latest version, ksh93t was recently released. Mr. Korn through his involvement with the Austin / SUS group has gone to great lengths to ensure correct POSIX behavior in ksh93. Ksh93 has shown itself to be much more efficient than other shells. And as someone has already mentioned, ksh93 is probably the closest thing to a defacto
shell as any other shell.

I understand that some people probably want mksh to be used in part because it is licensed under the BSD license. I also would like to see all software released with NetBSD under a BSD license (c'mon CLANG / LLVM!). But having said that, IMO mksh is not on par with ksh93 in performance and it differs from ksh93 in features. At some point in the future when mksh can show complete compatibility with ksh93 and achieve a close approximation of performance then I would think it appropriate to use mksh as the default shell in base.

Kent Wilson

Home | Main Index | Thread Index | Old Index