bug-inetutils
[Top][All Lists]
Advanced

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

[bug-inetutils] Shell capabilities. (Was: tests/syslogd.sh and /tmp)


From: Mats Erik Andersson
Subject: [bug-inetutils] Shell capabilities. (Was: tests/syslogd.sh and /tmp)
Date: Mon, 23 Jan 2012 01:25:15 +0100
User-agent: Mutt/1.5.18 (2008-05-17)

lördag den 21 januari 2012 klockan 08:11 skrev Alfred M. Szmidt detta:
>    >    BTW, syslogd.sh says #!/bin/sh, but it seems to use non-POSIX
>    >    extensions, no?
>    >
>    > What extentions do are you thinking of exactly?  I couldn't find
>    > anything particular
> 
>    I noticed the $((foo)) construct that, according to the Autoconf
>    manual, doesn't work on Solaris 10.  However I thought we had some
>    positive success on Solaris?
> 
> $((FOO)) is in IEEE 1003.1-2004, not sure about earlier.  Maybe
> Solaris uses a non-POSIX shell?

I wish I had access to Solaris 9 or 10 for building IU, but alas.

On OpenIndiana, sh(1) and most other shells are linked to different
personalities of ksh93(1), and so is the documentation.

On GNU/OpenSolaris, i.e., NexentaCore, sh(1) and some more are linked
to bash(1):

   /bin/sh -> bash

However, the manual page sh(1) claims to describe

   /usr/bin/sh        ( -> bash on NexentaCore )
   /usr/xpg4/bin/sh

of which the latter is a standalone variant of ksh(1).
The important observation is that the manual page sh(1)
is NOT mentioning either of $( ) or $(( )). Still tests
show that "/usr/xpg4/bin/sh" does understand both contructs
since it is closely related to the Korn shell.

My interpretation is that we should avoid "$( )" and "$(( ))"
if we want to support Solaris 10 or 9.

Regards,
  Mats



reply via email to

[Prev in Thread] Current Thread [Next in Thread]