Boost logo

Boost :

Subject: Re: [boost] [contract] toward N3351 concepts
From: Nevin Liber (nevin_at_[hidden])
Date: 2012-10-08 13:46:41


On 7 October 2012 19:29, Evgeny Panasyuk <evgeny.panasyuk_at_[hidden]> wrote:

> 3. Hardness of checking leads to situation when some compilers may check
> more cases than others. Even newer versions of same compiler can be smarter
> than predecessor. Does that mean that code which was compiled OK on older
> version of compiler will be broken on new, "smarter" one?
>

I'd say yes.

Is this really a different situation than a more aggressive optimizing
compiler breaking old code that happened to be dependent on undefined
behavior which just happened to work in the past?

-- 
 Nevin ":-)" Liber  <mailto:nevin_at_[hidden]>  (847) 691-1404

Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk