On 07/26/2012 11:38 AM, Jim Meyering wrote: > FYI. This bug seems hard enough to trigger and low enough > impact that the fix does not deserve an entry in NEWS. > Provoking the failure might be tricky, since in addition to > causing fstat failure, you'd have to make the intervening > close (via close_fd) clobber errno, probably via EINTR or EIO. > Not worth trying to write a test, IMHO.
Agreed on the test not being warranted. Good spot. cheers, Pádraig.