svn commit: r247960 - head/cddl/contrib/opensolaris/tools/ctf/cvt

[Ulrich Spörlein]

On Thu, 2013-03-07 at 20:45:20 -0500, Eitan Adler wrote:
> On 7 March 2013 18:00, Dimitry Andric <dim at freebsd.org> wrote:
> > On 2013-03-07 23:29, Andriy Gapon wrote:
> > ...
> >
> >> I was not really kidding when I said that this change, as is, is a nop:
> >
> > ...
> >
> >>> +                       bname = strrchr(match.iim_name, '/');
> >>> +                       bname = bname == NULL ? match.iim_name : bname +
> >>> 1;
> >>
> >>
> >> It would probably make sense to make use of bname after going to through
> >> all the
> >> trouble of calculating it:
> >>
> >>>                         match.iim_file = match.iim_name;
> >>
> >>
> >> Should be match.iim_file = bname ?
> >
> >
> > Yes, I actually had this in my test code, and committed the wrong diff
> > by accident. :(  Fixed in r247962, in any case.
> >
> >
> >
> >> Surprised that clang hasn't warned you about this one :-)
> >
> >
> > Clang currently does not have the -Wunused-but-set-variable warning; it
> > is apparently available in the static analyzer.  On the other hand, that
> > warning is one of the most annoying ones that newer gcc's have. :-)
> 
> I have found real bugs in FreeBSD's code base as a result of this
> warning, but the vast majority were more annoying than useful.
> 

That's why I've disabled them on
http://scan.freebsd.your.org/freebsd-head/

:)
Uli