[Bug c++/52964] New: No warning on negative array size in template instantatiation

[Bug c++/52964] New: No warning on negative array size in template instantatiation

[xinliangli at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

             Bug #: 52964
           Summary: No warning on negative array size in template
                    instantatiation
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: xinliangli@gmail.com


template <int N> struct S1 { int arr[N - 5]; };
template <typename T> struct S2 { S1<sizeof(T)> s1; };
template <typename T> void foo(T x) { S2<T> s2; }
void test() { foo(42); }

/* From Clang:
templates1.cpp:1:38: error: 'arr' declared as an array with a negative
size
template <int N> struct S1 { int arr[N - 5]; };
                                    ^~~~~
templates1.cpp:2:49: note: in instantiation of template class 'S1<4>'
requested here
template <typename T> struct S2 { S1<sizeof(T)> s1; };
                                               ^
templates1.cpp:3:45: note: in instantiation of template class 'S2<int>'
requested here
template <typename T> void foo(T x) { S2<T> s2; }
                                           ^
templates1.cpp:4:15: note: in instantiation of function template
specialization 'foo<int>' requested here
void test() { foo(42); }
==============================================     */

[Bug c++/52964] No warning on negative array size in template instantatiation

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-04-12 23:04:05 UTC ---
pinskia@server:~$ ~/treecombine-gcc/bin/gcc t.cc
t.cc: In instantiation of ‘struct S1<4>’:
t.cc:4:49:   required from ‘struct S2<int>’
t.cc:5:45:   required from ‘void foo(T) [with T = int]’
t.cc:6:21:   required from here
t.cc:3:43: error: size of array is negative

[Bug c++/52964] No warning on negative array size in template instantatiation

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

Manuel López-Ibáñez <manu at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |manu at gcc dot gnu.org

--- Comment #2 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2012-04-12 23:16:35 UTC ---
And with trunk we print:

pr52964.cc: In instantiation of ‘struct S1<4>’:
pr52964.cc:2:49:   required from ‘struct S2<int>’
pr52964.cc:3:45:   required from ‘void foo(T) [with T = int]’
pr52964.cc:4:21:   required from here
pr52964.cc:1:43: error: size of array is negative
 template <int N> struct S1 { int arr[N - 5]; };
                                           ^

We could easily print the caret for each line. And the locations could be much
better.

I also like more the order of Clang and the extra info about what each thing
is, but we would have to convince Gabriel and/or Jason.

David, what else do you see that could be improved?

[Bug c++/52964] No warning on negative array size in template instantatiation

[xinliangli at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

--- Comment #3 from davidxl <xinliangli at gmail dot com> 2012-04-12 23:19:33 UTC ---
(In reply to comment #2)
> And with trunk we print:
> 
> pr52964.cc: In instantiation of ‘struct S1<4>’:
> pr52964.cc:2:49:   required from ‘struct S2<int>’
> pr52964.cc:3:45:   required from ‘void foo(T) [with T = int]’
> pr52964.cc:4:21:   required from here
> pr52964.cc:1:43: error: size of array is negative
>  template <int N> struct S1 { int arr[N - 5]; };
>                                            ^
> 
> We could easily print the caret for each line. And the locations could be much
> better.
> 
> I also like more the order of Clang and the extra info about what each thing
> is, but we would have to convince Gabriel and/or Jason.
> 
> David, what else do you see that could be improved?

With -fsyntax-only, the warning is not emitted -- any reason for that?

David

[Bug c++/52964] No warning on negative array size in template instantatiation

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-04-12 23:23:31 UTC ---
(In reply to comment #3) 
> With -fsyntax-only, the warning is not emitted -- any reason for that?

Yes because -fsyntax-only does not do template instantatiations at all IIRC.

[Bug c++/52964] No warning on negative array size in template instantatiation

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

--- Comment #5 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2012-04-12 23:25:36 UTC ---
(In reply to comment #4)
> (In reply to comment #3) 
> > With -fsyntax-only, the warning is not emitted -- any reason for that?
> 
> Yes because -fsyntax-only does not do template instantatiations at all IIRC.

And is there a reason for that?

[Bug c++/52964] No warning on negative array size in template instantatiation

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

--- Comment #6 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-04-13 09:50:01 UTC ---
It's at least somewhat making -fsyntax-only less useful for C++ ... I'd use
-fsyntax-only to have all errors reported and thus all invalid CUs rejected ...

We do report non-syntax errors with -fsyntax-only after all.

[Bug c++/52964] No warning on negative array size in template instantatiation

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52964

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |DUPLICATE

--- Comment #7 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-16 23:08:42 UTC ---
Same issue, let's keep only one.

*** This bug has been marked as a duplicate of bug 54706 ***