svn commit: r329722 - head/cddl/usr.sbin/zdb

[Alan Somers]

On Sat, Feb 24, 2018 at 5:19 PM, Conrad Meyer <cem at freebsd.org> wrote:

> On Wed, Feb 21, 2018 at 8:40 PM, Alan Somers <asomers at freebsd.org> wrote:
> > On Wed, Feb 21, 2018 at 8:51 AM, Alan Somers <asomers at freebsd.org>
> wrote:
> >>
> >> Author: asomers
> >> Date: Wed Feb 21 15:51:48 2018
> >> New Revision: 329722
> >> URL: https://svnweb.freebsd.org/changeset/base/329722
> >>
> >> Log:
> >>   zdb: raise WARNS from 0 to 2
> >>
> >>   This has only been possible since r329694 and r329508
> >>
> >>   MFC after:    3 weeks
> >>   X-MFC-With:   329694
> >>   X-MFC-With:   329508
> >>   Sponsored by: Spectra Logic
> >>
> >> Modified:
> >>   head/cddl/usr.sbin/zdb/Makefile
> >>
> >
> > Looks like I broke the build with GCC-6.
>
> Yep, I can confirm it's broken.  I don't see a flag we can just turn
> off for the GCC warning, though, so the solution might have to be
> CFLAGS.gcc += -Wno-error.
>
> > I'll fix it, as soon as I can get
> > the toolchain setup.
>
> The easy way is just:
>
> $ pkg install amd64-xtoolchain-gcc
> $ make -sj$(sysctl -n hw.ncpu) buildworld CROSS_TOOLCHAIN=amd64-gcc
>
> Thanks,
> Conrad
>


There seem to be 3 problems: CTASSERT, assfail, and array bounds.  I've
already committed a fix for CTASSERT and I'm testing an assfail fix now.
The array bounds problem is harder, though.  The logic is complicated, but
I think it's sound.  I don't know why GCC is complaining, unless maybe it
thinks that an enum type is signed.  If I can't find a good solution, I'll
set -Wno-error as you suggest.

-Alan