r199053 - Clarify warn_cxx98_compat_attribute diagnostic

[Aaron Ballman]

After doing a bit more research and discussion off-list, I think
"generalized attribute" is acceptable.  So patch LGTM as-is.

Thanks!

~Aaron

On Sun, Jan 12, 2014 at 11:46 AM, Alp Toker <alp at nuanti.com> wrote:
>
> On 12/01/2014 15:29, Aaron Ballman wrote:
>
> On Sun, Jan 12, 2014 at 10:18 AM, Alp Toker <alp at nuanti.com> wrote:
>
> Author: alp
> Date: Sun Jan 12 09:18:06 2014
> New Revision: 199053
>
> URL: http://llvm.org/viewvc/llvm-project?rev=199053&view=rev
> Log:
> Clarify warn_cxx98_compat_attribute diagnostic
>
> Various attribute flavours are supported in C++98. Make it clear that this
> compatibility warning relates specifically to C++11-style generalized
> attributes.
>
> Modified:
>     cfe/trunk/include/clang/Basic/DiagnosticParseKinds.td
>     cfe/trunk/test/SemaCXX/cxx98-compat.cpp
>
> Modified: cfe/trunk/include/clang/Basic/DiagnosticParseKinds.td
> URL:
> http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/DiagnosticParseKinds.td?rev=199053&r1=199052&r2=199053&view=diff
> ==============================================================================
> --- cfe/trunk/include/clang/Basic/DiagnosticParseKinds.td (original)
> +++ cfe/trunk/include/clang/Basic/DiagnosticParseKinds.td Sun Jan 12
> 09:18:06 2014
> @@ -517,7 +517,7 @@ def warn_cxx98_compat_nullptr : Warning<
>  def warn_cxx98_compat_alignas : Warning<"'alignas' is incompatible with
> C++98">,
>    InGroup<CXX98Compat>, DefaultIgnore;
>  def warn_cxx98_compat_attribute : Warning<
> -  "attributes are incompatible with C++98">,
> +  "generalized attributes are incompatible with C++98">,
>
> What are "generalized" attributes? I think it would be better-worded
> as "C++11 attributes are incompatible with C++98" (this is more
> consistent with other parser diagnostics, as well).
>
>
>
> Hi Aaron,
>
> I'd say it's time to use the more natural name for C++11 attributes because
> the syntax now appears in a published standard. That's the usual lifecycle
> for clang language features moving from experimental to fully supported.
>
> As for the name itself, it's the terminology adopted by the C++ community
> and also the name by which we advertise the feature on our own C++ status
> page at http://clang.llvm.org/cxx_status.html:
>
>   Generalized attributes    N2761    Clang 3.3 (1)
>
> There's also a compelling reason _not_ to keep the old experimental naming
> scheme indefinitely -- doing so will cause terminology dissonance as other
> language dialects like ISO C look to adopt the new syntax, likewise if we
> decide to introduce generalized attributes as a clang extension to C11 in
> the meantime.
>
> Alp.
>
>
>
> Otherwise, LGTM!
>
> ~Aaron
>
>
> --
> http://www.nuanti.com
> the browser experts