Re: [PATCH 1/5] Poison non-POD memset & non-trivially-copyable memcpy/memmove

[Pedro Alves <palves@redhat.com>]

On 04/25/2017 09:24 AM, Yao Qi wrote:
> Pedro Alves <palves@redhat.com> writes:
> 
> Hi Pedro,
> 
>> +/* True if "T *" is relocatable.  I.e., copyable with memcpy/memmove.
>> +   I.e., T is either trivially copyable, or void.  */
>> +template<typename T>
>> +struct IsRelocatable
>> +  : gdb::Or<std::is_void<T>,
>> +	    std::is_trivially_copyable<T>>
>> +{};
> 
> This breaks the build with gcc 4.8,
> 
> In file included from ../../binutils-gdb/gdb/common/common-defs.h:85:0,
>                  from ../../binutils-gdb/gdb/defs.h:28,
>                  from ../../binutils-gdb/gdb/gdb.c:19:
> ../../binutils-gdb/gdb/common/poison.h:66:6: error: ‘is_trivially_copyable’ is not a member of ‘std’
>       std::is_trivially_copyable<T>>
>       ^
> 

Sorry, I thought I had tested gcc 4.8, but clearly I did not.  I'll fix it
as soon as I have a chance, probably by disabling the poisoning on
older compilers.

> you probably have already received the buildbot fail message
> https://gdb-build.sergiodj.net/builders/Ubuntu-AArch64-m64/builds/1845
> 
> is_trivially_copyable is missing on 4.9 too,
> https://gcc.gnu.org/onlinedocs/gcc-4.9.4/libstdc++/manual/manual/status.html#status.iso.2011
> and it was added in gcc 5, as far as I know.

Thanks,
Pedro Alves