[cfe-commits] r150625 - in /cfe/trunk: include/clang/Basic/DiagnosticSemaKinds.td include/clang/Sema/Initialization.h lib/Sema/SemaInit.cpp test/Misc/warning-flags.c test/SemaCXX/cxx0x-initializer-aggregates.cpp

[Eli Friedman]

On Wed, Feb 15, 2012 at 3:39 PM, Richard Smith <richard at metafoo.co.uk> wrote:
> On Wed, Feb 15, 2012 at 2:49 PM, Eli Friedman <eli.friedman at gmail.com>
> wrote:
>>
>> On Wed, Feb 15, 2012 at 2:38 PM, Richard Smith
>> <richard-llvm at metafoo.co.uk> wrote:
>> > Author: rsmith
>> > Date: Wed Feb 15 16:38:09 2012
>> > New Revision: 150625
>> >
>> > URL: http://llvm.org/viewvc/llvm-project?rev=150625&view=rev
>> > Log:
>> > Support GCC's bug^Wextension allowing class array members to be
>> > initalized by a
>> > parenthesized braced-init-list in the base/member initialization list.
>> >
>> > Modified:
>> >    cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td
>> >    cfe/trunk/include/clang/Sema/Initialization.h
>> >    cfe/trunk/lib/Sema/SemaInit.cpp
>> >    cfe/trunk/test/Misc/warning-flags.c
>> >    cfe/trunk/test/SemaCXX/cxx0x-initializer-aggregates.cpp
>> >
>> > Modified: cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td
>> > URL:
>> > http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td?rev=150625&r1=150624&r2=150625&view=diff
>> >
>> > ==============================================================================
>> > --- cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td (original)
>> > +++ cfe/trunk/include/clang/Basic/DiagnosticSemaKinds.td Wed Feb 15
>> > 16:38:09 2012
>> > @@ -3550,7 +3550,10 @@
>> >   "cannot initialize array of type %0 with non-constant array of type
>> > %1">;
>> >  def ext_array_init_copy : Extension<
>> >   "initialization of an array of type %0 from a compound literal of type
>> > %1 is "
>> > -  "a GNU extension">;
>> > +  "a GNU extension">, InGroup<GNU>;
>> > +def ext_array_init_parens : ExtWarn<
>> > +  "parenthesized initialization of a member array is a GNU extension">,
>> > +  InGroup<GNU>;
>>
>> Can you put this in a more specific warning group, and make it
>> DefaultError?  Since only very new gcc versions accept this, we really
>> want to discourage anyone from using it.
>
>
> Sure thing. Do you want it in a subgroup of -Wgnu, or do you want for
> -Wno-gnu to leave this enabled?

-Wno-gnu should probably leave this enabled.  Assuming this issue has
been reported against gcc, I think they're likely to fix it rather
than leave it in as an extension.

-Eli