Boost logo

Boost :

Subject: Re: [boost] C++03 / C++11 compatibility question for compiled libraries
From: degski (degski_at_[hidden])
Date: 2018-02-12 20:16:59


On 12 February 2018 at 11:23, Paul A. Bristow via Boost <
boost_at_[hidden]> wrote:

> Some on this thread seem to view 'support' and 'validated' as binary
> features.
>

I still think they are. If you cannot use the tools created by the library
author to validate that the library functions well over a set of tests
deemed relevant by that library author, one has to conclude that that
C++-Version is not supported.

> Boost (and life in general) really isn't as simple as that.

Test coverage is not ever going to be 100%.

In principle I couldn't care less as to what C++-standard a library is
written in. I wouldn't even mind if boost had a number pure C-Libs. BUT, a
library, f.e. not supporting C++11 moves, starts to violate the principle
that "I don't pay for what I don't use.", because stuff gets copied, while
it could be moved. Variadic templates is equally not to be sniffed upon.

All this looking backwards, let's not disenchant those people who insist on
using an old compiler, where does that come from. If and when you're not
moving to C++11 you are living in the dark ages. Yes, at some point it was
*new*, and there might be bugs (luckily there were no bugs in older boost
versions). In the meanwhile C++11 is old news.

One can only speculate as to why a number of people involved with boost are
soooooooo stuck in the past, but as in many issues/problems in the world in
general, the first thing to do is, IMHO, follow the money. That's an old
lesson I learned, long time ago.

degski


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