On Sat, Feb 23, 2008 at 11:06:52AM +0100, Ralf Wildenhues wrote: > * Ralf Wildenhues wrote on Fri, Feb 22, 2008 at 06:47:54PM CET: > > > > 2) NetBSD ksh has a bug. At least I *think* that is a bug, I haven't > > checked SUSv3 yet. This: > > foo=foo; t='[ ]'; case " foo bar " in *$t$foo$t*) echo yes;; esac
It seems NetBSD ksh ( == @(#)PD KSH v5.2.14 99/07/13.2) has lots of bugs. (I should have remembered make uses sh which works, but gmake/autotools choose bash over ksh over sh (is that right?)) quartz% sh $ foo=foo; t='[ ]'; case " foo bar " in *$t$foo$t*) echo yes;; esac yes $ exit quartz% ksh $ echo $KSH_VERSION @(#)PD KSH v5.2.14 99/07/13.2 $ set -o posix $ foo=foo; t='[ ]'; case " foo bar " in *$t$foo$t*) echo yes;; esac $ set +o posix $ foo=foo; t='[ ]'; case " foo bar " in *$t$foo$t*) echo yes;; esac $ exit I think I am sounding like a broken record! PR/26493 http://mail-index.netbsd.org/current-users/2004/07/27/0013.html http://mail-index.netbsd.org/current-users/2004/07/27/0014.html http://mail-index.netbsd.org/current-users/2004/07/28/0004.html http://mail-index.netbsd.org/current-users/2004/07/30/0006.html http://mail-index.netbsd.org/current-users/2004/08/31/0008.html http://mail-index.netbsd.org/tech-pkg/2005/02/18/0000.html http://mail-index.netbsd.org/current-users/2005/05/13/0016.html http://mail-index.netbsd.org/current-users/2006/09/13/0008.html usw. > > should IMHO print yes, but doesn't with this shell. > > This violates POSIX. Can you please file a bug report with NetBSD? I'll add another one to the list. > I've applied this patch to master, and likewise (without the > keep_testdirs bits) to branch-1-10. For now I don't know of > a better way to automatically test for this issue than to > encourage users to try MAKE=gmake. ... or pick sh over (pd)ksh? Cheers, Patrick