On 03/ 2/10 01:44 PM, Peter Tribble wrote:
On Tue, Mar 2, 2010 at 3:47 PM, Garrett D'Amore<[email protected]>  wrote:
On 03/ 2/10 03:57 AM, [email protected] wrote:
On Tue, Mar 2, 2010 at 7:02 AM, Garrett D'Amore<[email protected]>    wrote:

I'm just thinking, isn't it time we just bit the bullet and told our
users
to start using the POSIX defaults, instead of continuing to supply the
legacy compatibility stuff forever.

I believe a number of scripts requires certain tools from that package so
I think we always need to install them; however, I think it is better
that we try our level to merge them into the standard location.

And *that* is what I was saying.  I think we should finally adopt the POSIX
semantics.
Provided it's a reasonable merge. For example, I updated du
(seems like ages ago now) so that the regular and xpg4 versions
both understood -d and -x. There are going to be some differences
that are unavoidable, but allowing the use of common old-fashioned
flags where there aren't conflicts makes transition so much easier.

I haven't looked recently to see how big a problem this is still. What I
do know is that working in a heterogeneous environment (and with
different configurations mixed in) the minor differences in behaviour
drive me up the wall.

The point here is that in OpenSolaris/Solaris Next, maybe we can finally ditch the legacy Sun semantics. I'd only do that where we couldn't *adopt* the Sun semantics as is. For example, command line switches only understood by the non-XPG4 version could be incorporated into the merged version.

The tricky part is dealing with the subtle semantic differences. For example, what does date +%C output? For /usr/bin, its the default strftime output. For standard conforming date it is the century number. (Which means you should never use %C in a date format string in a portable shell script! :-) Its these kinds of legacy Sun semantics that I think we can finally bite the bullet and eliminate, offering the POSIX conformant semantics instead.

    -- Garrett
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to