This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: fnmatch() behaves oddly with *s and FNM_LEADING_DIR
Ulrich Drepper <drepper@redhat.com> writes:
> Paul Eggert <eggert@twinsun.com> writes:
> > The glibc 2.1.3 manual documents FNM_LEADING_DIR to behave the way
> > that GNU tar uses it. Has it been withdrawn recently? If so, why?
> > GNU tar has been using that flag ever since it was added to glibc in 1992.
>
> Tar is probably using it the way it is intended. But look at the
> implementation and you'll see that it handles only a few cases.
Sorry, I don't follow. From tar's point of view, it would be OK if
fnmatch misbehaved only in areas that tar doesn't exercise. But
Debian bug 59829 came from someone who used tar, and found that tar
misbehaves because fnmatch misbehaves.
> Making it general usable (which is not necessary in the first place
> IMO) only slows down normal use.
Why isn't it necessary to make the feature work in general?