On 01/28/13 08:40, Pavel Raiskup wrote: > + if (u > st->stat.st_size) > + /* Do *not* overwrite stat.st_size when it has been set by > + GNU.sparse.realsize (func sparse_size_decoder) already. */ > + st->stat.st_size = u;
Hmm, this doesn't feel right. Can't the sparse size be greater than the real size, in some (unusual) cases? We should overwrite the size when realsize has not been specified, not just when our size exceeds what's in st_size.
