On Tue, 18 Jan 2005, Gunnar Ritter wrote:
> > As best I can tell, an option which tells ls which filenames are to be
> > sent to output is *independent* of whether or not ls's output formatting
> > deals (somehow) with the embedded-newline problem.
>
> If the fix should become a recommendation against using ls for scripts,
> there is no point in adding -A, so no, they are not independent.
Even a strictly-interactive ls would benefit from -A. Moreover, even if
the best fix for the embedded-newline problem is a completely new facility
rather than a safe-output option to ls, the issue will come up again in
connection with the new facility (since it will want to copy many of ls's
behaviors and options to simplify script conversion). So there's no reason
to delay consideration of -A pending consideration of the embedded-newline
problem; the two *are* effectively independent.
If postponing -A would stop everyone from making unsafe use of ls in
scripts, it might have some merit. But it won't.
Henry Spencer
yyyyy@xxxxxxxxxxxxx
|