Re: [PATCH][RFC] Add a subset of -Warray-bounds warnings to C/C++ front ends

["H.J. Lu" <hjl.tools@gmail.com>]

Hi Simon,

In http://gcc.gnu.org/ml/gcc-patches/2008-04/msg01945.html, you said

---
A side effect of copying these warnings up into the language frontends is
that warnings are now printed even if the array access is in dead or
inaccessible code.
---

So this behavior is intentional. I don't see how it will be useful. Should
this patch be reverted for now?

Thanks.


H.J.
---
On Fri, May 2, 2008 at 5:50 AM, H.J. Lu <hjl.tools@gmail.com> wrote:
> Hi Simon,
>
>  Your patch breaks gcc bootstrap:
>
>  http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36108
>
>
>  H.J.
>
>
> On Thu, May 1, 2008 at 12:07 PM, Simon Baldwin <simonb@google.com> wrote:
>  > Thanks all.  Committed, with suggested doc/invoke.texi change, as revision
>  > 134865.
>  >
>  >
>  >
>  >
>  >  Mark Mitchell wrote:
>  >
>  > > Simon Baldwin wrote:
>  > >
>  > > > Thanks to a well targeted pointer from Tom Tromey, I've now added the
>  > extra logic to suppress warnings about "sizeof(a[-1])" in the C and C++
>  > frontends.  Attached is the revised version.
>  > > >
>  > >
>  > > Thanks for following up.  I think this is a worthwhile patch.
>  > >
>  > >
>  > > > +This option performs a subset of checks in unoptimized compilations,
>  > and
>  > > > +stricter checking when @option{-ftree-vrp} is active
>  > > >  (default for -O2 and above).
>  > > >
>  > >
>  > > I think "stricter" could be made more explicit.  How about something like:
>  > >
>  > > This option detects some cases of out-of-bounds accesses in unoptimized
>  > compilations.  More cases are detected when @option{-ftree-vrp} is enabled.
>  > (The @option{-ftree-vrp} option is enabled automatically when compiling with
>  > @option{-O2} or higher optimization options.)
>  > >
>  > > Joseph, do you have any comments on the C changes?  Simon, if you do not
>  > hear otherwise from Joseph within 72 hours, this patch is OK.
>  > >
>  > > Thanks,
>  > >
>  > >
>  >
>  >
>