On 1/4/21 10:30 PM, Simon Marchi wrote:
> On 2021-01-04 3:57 p.m., Bernd Edlinger wrote:
>> Hi,
>>
>> with Luis' commit of today the trunk is no longer able to
>> be compiled with gcc-4.x.
>>
>> The problem is std::is_trivially_default_constructible is
>> not defined before gcc-5 although the compiler supports C++11 
>>
>> I am not sure about what's the best approach for conditionally
>> enabling the code, especially for compilers other than g++.
>>
>>
>> This fixes the build for me.
>> Is it OK for trunk?
>>
>>
>> Thanks
>> Bernd.
>>
> 
> We have other instances of this, I'd suggest inspiring yourself from:
> 
> https://sourceware.org/git/?p=binutils-gdb.git;a=blob;f=gdb/unittests/enum-flags-selftests.c;h=e3b6cf81d07f1eaefec72bce131d1c75ce00ef82;hb=HEAD#l66
> 
> These defines are defined here:
> 
> https://sourceware.org/git/?p=binutils-gdb.git;a=blob;f=gdbsupport/traits.h;h=f545edbb0d93e95f65e954fbf54cbc8843e5239a;hb=HEAD#l28
> 
> As you see, we already have a HAVE_IS_TRIVIALLY_CONSTRUCTIBLE, we could
> use it.  The code in trad-frame.c could be changed to use
> std::is_trivially_constructible instead of
> std::is_trivially_default_constructible, I believe it's the same when
> passing no Args... to is_trivially_constructible.  See "Possible
> implementation" in:
> 
> https://en.cppreference.com/w/cpp/types/is_default_constructible
> 
> Simon
> 

Okay, good point.

I updated the patch following your suggestion.

Is the updated patch OK?


Thanks
Bernd.