[ksh93-integration-discuss] printf syntax and /sbin/shcompatibility

[Roland Mainz]

Shawn Walker wrote:
> On Feb 7, 2008 11:13 PM, David Korn <dgk at research.att.com> wrote:
> > cc: swalker at opensolaris.org
> > Subject: Re: [ksh93-integration-discuss] printf syntax and /sbin/sh compatibility
> > > I just ran into an odd behaviour tonight (see bug
> > > http://defect.opensolaris.org/bz/show_bug.cgi?id=502).
> > >
> > > The following script:
> > >
> > > printf "\n----------------------\n"
> > > printf "   IP configuration   \n"
> > > printf "----------------------\n\n"
> > >
> > > ...works under ye olde /sbin/sh (bourne shell), but fails with
> > > "unknown option" under ksh93.
> > >
> > > Since I'm rather daft about some nuances of shell syntax, can someone
> > > explain this to me?
[snip]
> >
> > This is not a defect.  If the first argument a command begins with a -
> > it is treated as an option, not an  argument.  The operand -- indicates
> > end of options.
> >
> > Change
> >         printf "----------------------\n\n"
> > to
> >         printf -- "----------------------\n\n"
> 
> In this case, it isn't a script I wrote.
> 
> It's part of a shell script that is used during a driver install for a
> Marvell Yukon Gigabit Ehternet Adapter.
> 
> Sadly, Solaris' version of bourne shell? /sbin/sh allows the "bad
> syntax" I mentioned above.

Erm, the issue is not related to /sbin/sh. The Bourne shell (= /sbin/sh)
just has no built-in "printf" command and uses /usr/bin/printf

> I wonder if some sort of compatibility mode can be achieved for
> bourne's printf syntax.

Erm... $ man -s1 printf # says:
-- snip --
ATTRIBUTES
 See attributes(5) for descriptions of the  following  attri-
 butes:

 ________________________________________________________
|       ATTRIBUTE TYPE    |       ATTRIBUTE VALUE       |
|_________________________|_____________________________|
| Availability            | SUNWloc                     |
|_________________________|_____________________________|
| CSI                     | enabled                     |
|_________________________|_____________________________|
| Interface Stability     | Standard                    |
|_________________________|_____________________________|
-- snip --

I'll try some analysis (Don Cragun may correct me):
The "interface stabilty" is defined as "Standard", e.g. /usr/bin/printf
should be "standards conforming". Since
http://www.opengroup.org/onlinepubs/009695399/utilities/printf.html does
not list any exception in argument handling for the "printf" utilty the
normal rules for argument parsing apply (see
http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap01.html)
- which require that "printf" just be able to handle options and treat
arguments which start with a '-' character as options (if they are
before an argument with the value "--").
AFAIK the current behaviour of Solaris's /usr/bin/printf is therefore a
bug (and per Knut Reiner's comment in
http://defect.opensolaris.org/bz/show_bug.cgi?id=502 the POSIX test
suites complain about the issue).

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) roland.mainz at nrubsig.org
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 7950090
 (;O/ \/ \O;)