Re: [boost] The future and present of Boost

Subject: Re: [boost] The future and present of Boost
From: Edward Diener (eldiener_at_[hidden])
Date: 2018-10-24 22:05:38

On 10/24/2018 5:21 PM, Peter Dimov via Boost wrote:
> Edward Diener wrote:
>> I don't get this. Why can't library X, which needs C++11, depend on 
>> library Y, which needs C++03 but works fine if compiled with C++11. 
>> What am I missing here ?
> 
> I don't know how to answer this, and I don't know how it relates to what 
> I was saying.
> 
> What I was saying is this: the user, or someone acting on his behalf 
> such as a package maintainer, types f.ex.
> 
>  Â Â  b2 toolset=clang install
> 
> and obtains a prebuilt Boost.
> 
> Then the user types f.ex.
> 
>  Â Â  clang++ -std=c++17 myapp.cpp -lboost_pumpkin
> 
> and expects it to work. If boost_pumpkin autodetects C++11/C++14/C++17 
> and changes its API to match, it doesn't work.

Is not this a user problem, irrelevant of cxxstd level ? The user is 
given a Boost package compiled  with the default cxxstd level for clang 
and then attempts to use it imposing his own cxxstd level. I would 
expect disaster in any case except when the default cxxstd level matches 
what the user specifically uses, regardless of what those levels ( 
c++03, c++11, c++14, c++17, c++20 ) actually are. I can not imagine how 
this can be a Boost ( or any library ) problem as opposed to a user problem.

> 
> It doesn't because step 1 builds libboost_pumpkin.a with the default 
> -std=c++03.
> 
> Same with g++ 5.
> 
> Not same with g++ 6 and above, whose default is C++14, but still 
> possible if the library adapts to C++17 by using f.ex. <filesystem>.