On Mon, Jan 28, 2013 at 01:16:37PM +0100, Michal Židek wrote: > On 01/28/2013 10:44 AM, Jakub Hrozek wrote: > >On Mon, Jan 28, 2013 at 10:37:19AM +0100, Pavel Březina wrote: > >>We should also test the return value of close() on the succeeding > >>lines. I wonder why Coverity didn't argue here. > > > >My guess is that we also don't check close() return value elsewhere. The > >way I've seen Coverity work is that if most other occurences are > >checked, then Coverity raises a flag is one is not. > > Yes I though so too. That is also the reason, why I didn't check the > return value of close here. If we want to check it, we must change > it everywhere where close() is used (Otherwise Coverity will > probably complain). Do we want to check close() return values? If so > I can post another patch related to close() only that will check it > everywhere, but I think it is not necessary. >
No, I don't think mass-fixing the close calls is necessary. In new code, we should try to check the return value and warn to DEBUG logs, but usually there's not much we can do anyway. _______________________________________________ sssd-devel mailing list [email protected] https://lists.fedorahosted.org/mailman/listinfo/sssd-devel
