[PATCH 3/9] posix: Allow glob to match dangling symlinks [BZ #866]

[Paul Eggert]

On 09/11/2017 07:38 AM, Zack Weinberg wrote:
> If this is done, the new symbol version should attempt to validate the
> altdirfuncs structure on entry

I don't see how glob could do that reliably and cheaply. How does one 
validate that a pointer to a stat-like function is actually a pointer to 
a stat-like function? All one can do is call the function and pray.