[systemd-devel] [systemd-commits] autogen.sh src/core src/shared src/tmpfiles

[Zbigniew Jędrzejewski-Szmek]

On Mon, Oct 21, 2013 at 07:56:12PM +0200, Kay Sievers wrote:
> On Mon, Oct 21, 2013 at 7:49 PM, Zbigniew Jędrzejewski-Szmek
> <zbyszek at in.waw.pl> wrote:
> > On Mon, Oct 21, 2013 at 06:55:12PM +0200, Kay Sievers wrote:
> >> On Mon, Oct 21, 2013 at 6:33 PM, Kay Sievers <kay at vrfy.org> wrote:
> >> > On Mon, Oct 21, 2013 at 6:11 PM, Zbigniew Jędrzejewski-Szmek
> >> > <zbyszek at in.waw.pl> wrote:
> >> >
> >> >>> Hmm, I'm running up-to-date F20, from today here. And have no issues.
> >> >>> Tried a "git clean"?
> >> >>
> >> >> I now did a 'yum update', and with a bunch of packages installed, most
> >> >> notably gcc-4.8.2, this problem is gone. It seems that we're on the
> >> >> bleeding edge regarding gcc too.
> >> >
> >> > It was the combination with FORTIFY_SOURCE, I disabled that again for
> >> > "debug builds", and pushed it. No errors are thrown any more. We can
> >> > enable FORTIFY_SOURCE for debug build too when things have catched up a bit.
> >>
> >> It builds without a warning on F19 and not updated F20 now,
> >> but still catches this serious bug, which was not detected
> >> before the -Og change:
> >>   http://cgit.freedesktop.org/systemd/systemd/commit/?id=bd4b5df2cb253f130ecdac604e865e3f832c316a
> >
> > (If -Og works everywhere, this might be a moot point, but...)
> > we (as in developers) can compile with whatever settings work the best
> > for us (e.g. with gcc + clang + clang-static-analyzer), but I don't
> > think enabling compiler warnings makes much sense for users.
> 
> Not sure what you mean, autogen.sh always recommended/printed -O0 by
> default, and "./autogen c" used it too, now it uses -Og instead. But
> sure, people can still pass whatever they think is the best.
What I meant is that we don't need to encourage flags which are good
for developers as the default. Current recommendations are OK.

Zbyszek