On Mon, Dec 30, 2013 at 04:58:50PM +0100, Mike Belopuhov wrote: > On 30 December 2013 16:35, Loganaden Velvindron <lo...@elandsys.com> wrote: > > On Mon, Dec 30, 2013 at 08:42:00AM -0500, Ted Unangst wrote: > >> On Mon, Dec 30, 2013 at 13:53, Mike Belopuhov wrote: > >> > On Mon, Dec 30, 2013 at 03:59 -0800, Loganaden Velvindron wrote: > >> >> On Mon, Dec 30, 2013 at 12:45:47PM +0100, Mike Belopuhov wrote: > >> >> > On Sun, Dec 29, 2013 at 22:45 -0800, Loganaden Velvindron wrote: > >> >> > > Hi All, > >> >> > > > >> >> > > From NetBSD: > >> >> > > > >> >> > > Plug memory leak. Coverity CID 1596 > >> >> > > > >> >> > > >> >> > memory leak? can you please elaborate where else this memory > >> >> > is "leaking" if not back to the system. > >> >> > >> >> After a short discussion on IRC, and some digging, it makes sense > >> >> that the system cleanups up automatically, and therefore it is not > >> >> necessary. > >> >> > >> > > >> > the patch can be committed though for the sake of better style. > >> > >> but please review to make sure they are correct. > > > > It would be great if the ldconfig, and the various maintainers could > > review the diff, and help me improve the diffs, or discard them. > > > > but the same applies to the ldconfig "fix". buildhints is called right > before exit. there are no leaks there. the only thing that can be > improved is unlinking tmpfilenam.
Maybe so, but ever since I've known of OpenBSD, it has always "preached" good coding practices and exemplified through its base source code. So why are you discouraging such fixes? You are almost saying that any non-daemon program should not bother with free(), close() and similar resource de-allocation function, because the system will do the reclaiming after program exits. That's rubbish. Regards, --patrick > > I'm mostly interested in finding the small security issues and fixing > > them, rather than fixing style issues :-) > > >
