Re: [C++ Patch] PR 36760 (again)

[Paolo Carlini <paolo dot carlini at oracle dot com>]

Hi Mark,
> This is OK for mainline.
OK, thanks.
>   I don't see a regression justifying 4.3. 
Yeah, it's just that the warning from <functional> is rather annoying 
and, back in 4.2.x times, the equivalent functionalities in 
<tr1/functional> didn't trigger any warning... But it's of course your 
call, let's fix the issue in mainline, then maybe I can adjust the 
library for 4_3.
> Also, please add a test case specifically for this patch, if not 
> already present in the testsuite, or enhance one already there if 
> there's already a test for this warning.  I think all three cases:
>
>   const int f();
>
>   template <typename T>
>   const int g(T);
>
>   // No warning.
>   template <typename T>
>   const T h();
>
> should appear.  Test case change pre-approved.
I see your point and was indeed prepared to work on a new testcase but 
luckily all those cases are *already* covered in Wreturn-type-4.C.

Thanks again,
Paolo.