Sync from the stable (b5) branch of coreutils:

Sync from the stable (b5) branch of coreutils:

(fts_children): Don't let close() clobber errno from
failed fchdir().

(fts_stat): When following a symlink-to-directory,
don't necessarily interpret stat-fails+lstat-succeeds as indicating
a dangling symlink.  That can also happen at least for ELOOP.
The fix: return FTS_SLNONE only when the stat errno is ENOENT.
FYI, this bug predates the inclusion of fts.c in coreutils.

(fts_open): Put new maxarglen declaration and uses
in their own block, so pre-c99 compilers don't object.

Avoid the double-free (first in fts_read, second in fts_close) that
would occur when an `active' directory is made inaccessible (e.g.,
via chmod a-x) during a traversal.

(fts_read): After a failed fchdir, update sp->fts_cur
before returning.  Reproduce this failure by
mkdir -p a/b; cd a; chmod a-x . b
Reported by Stavros Passas.