Re: [boost] Boost.GSL?

Subject: Re: [boost] Boost.GSL?
From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2015-10-01 03:14:08

Beman Dawes <bdawes <at> acm.org> writes:

> The full talk is available at https://www.youtube.com/watch?
v=1OEu9C51K2A
> 
> Herb Sutter's take the next day is also essential to understanding the
> whole picture. See https://www.youtube.com/watch?v=hEx5DNLWGgA

I finally got through the second talk. It was really... disappointing. It 
felt like cheap circus for little kids. We all know that some logic 
deductions can be made in a trivial cases. I did not see any adult 
discussion about why they believe these examples are representative. IMO 
99.999% of the real use case will not like this branch-less, loop-less 
code. Real code will look more like:

void foo( int i )
{
   vector<int> v{1};
   int& ref = v[0];

   if( i%2 )
     v.push_back( 2 );

   if( i%2 )
     i = ref;
}

What is their local static analysis suggest us to do in this case? If code 
is wrong, how should we write it in modern C++?

I do not want to discourage this effort by no means, and I really believe 
we can use modern C++ to produce much better code. Yet, this talk did not 
lift my spirit.

Gennadiy