Re: [coreutils] Re: Builds of coreutils-8.6.19-1369f

[Ralf Wildenhues]

* Jim Meyering wrote on Sun, Nov 07, 2010 at 12:22:13PM CET:
> Ralf Wildenhues wrote:
> > * Jim Meyering wrote on Sat, Nov 06, 2010 at 06:14:05PM CET:
> >> Ralf Wildenhues wrote:
> >> > - Posix does not require IFS to be set upon script startup.  It only 
> >> > requires
> >> > that, if IFS is unset, word splitting ought to behave as if it were set 
> >> > to
> >> > space, tab, and newline, in that order.  That is why the startup code of 
> >> > all
> >> > autotools scripts explicitly sets IFS, so that saving and restoring 
> >> > works.
> >>
> >> Thanks.  That's good to know, but it appears not to be an issue
> >> for any shell that gets past the init.sh tests.
> >
> > Are you sure?  It affects any test that saves and restores IFS in any
> > way.

> Do you know of a modern shell that acts as you describe?
> I've run the following on a decent selection of systems,
> and saw no failure:

Some version of dash did that, IIRC.  A wild guess would
be RH 6 (yes, quite a while ago, I haven't come across it
again since).  Point is that Posix allows the behavior.

> If some shell does fail, perhaps init.sh should use a patch like the 
> following.
> I hesitate to add code like this, unless there's at least one
> known system where it makes a difference.

The fix looks good to me.

Thanks,
Ralf

> --- a/tests/init.sh
> +++ b/tests/init.sh
> @@ -330,6 +330,12 @@ setup_()
>      || fail_ "failed to create temporary directory in $initial_cwd_"
>    cd "$test_dir_"
> 
> +  # As autoconf-generated configure scripts do, ensure that IFS is defined
> +  # initially, so that saving and restoring $IFS works.
> +  gl_init_sh_nl_='
> +'
> +  IFS=" ""   $gl_init_sh_nl_"
> +
>    # This trap statement, along with a trap on 0 below, ensure that the
>    # temporary directory, $test_dir_, is removed upon exit as well as
>    # upon receipt of any of the listed signals.
>