Status of building problem with GCC 6

[Kim Barrett]

> On Jul 2, 2016, at 5:18 AM, Andrew Haley <aph at redhat.com> wrote:
> 
> On 01/07/16 20:13, Kim Barrett wrote:
>> -fno-lifetime-dse is a relatively recent option, and needs to be
>> conditionalized.
> 
> I think that we don't need -fno-lifetime-dse any more because the
> bug which needed it was fixed by 8034812.

My understanding was that -fno-lifetime-dse was needed to prevent
elimination of code like that in the #ifdef ASSERT block in
Node::operator new, e.g. treating storage as an object before the
constructor has been called.  There was also some non-ASSERT
code there that was removed by 8034812.  And as I mentioned in
8160742, I’m pretty sure I’ve seen other occurrences of this “pattern”.