[Libreoffice-commits] .: sal/inc

[Lubos Lunak]

On Sunday 09 of December 2012, Libreoffice Gerrit user wrote:
>  sal/inc/rtl/ustrbuf.hxx |    7 +++++++
>  1 file changed, 7 insertions(+)
>
> New commits:
> commit dc6277ef41dcf529760ad894e0aa124f18722c33
> Author: Tor Lillqvist <tml at iki.fi>
> Date:   Sun Dec 9 22:42:04 2012 +0200
>
>     WaE: deleted function definitions are a C++11 extension
>
>     Change-Id: I8f070fb53c0310761c3a0a2407d48ca1c1ac5e18
>
> diff --git a/sal/inc/rtl/ustrbuf.hxx b/sal/inc/rtl/ustrbuf.hxx
> index 3814a6f..9850bfa 100644
> --- a/sal/inc/rtl/ustrbuf.hxx
> +++ b/sal/inc/rtl/ustrbuf.hxx
> @@ -579,10 +579,17 @@ public:
>      // Pointer can be automatically converted to bool, which is unwanted
> here. // Explicitly delete all pointer append() overloads to prevent this
> // (except for char* and sal_Unicode* overloads, which are handled
> elsewhere). 
> +#if defined(__clang__) && defined(LIBO_WERROR) 
> +#pragma GCC diagnostic push
> +#pragma GCC diagnostic ignored "-Wc++11-extensions"
> +#endif
>      template< typename T >
>      typename internal::Enable< void,
>          !internal::CharPtrDetector< T* >::ok &&
> !internal::SalUnicodePtrDetector< T* >::ok >::Type append( T* ) = delete;
> +#if defined(__clang__) && defined(LIBO_WERROR)
> +#pragma GCC diagnostic pop
> +#endif
>  #endif
>  #endif

 That doesn't make sense. This code is inside #ifdef that gets enabled only 
inside configure part that requires C++11 support, so this part should get 
compiled only if we compile with C++11 support, thus there should be no way 
to get this warning. So how can anybody possibly need this?

 And, not to put a too fine point on it, I'd appreciate it if you could try 
something a bit more systematic such as fixing the configure check that's 
been put there for a reason instead of a half-assed band-aid somewhere in the 
code. If you think that's too much work, maybe you'll change your mind as 
you'll have to do this for every place where I add this next.

-- 
 Lubos Lunak
 l.lunak at suse.cz