Re: [boost] [cmake] Minimum viable cmakeification for Boost

Subject: Re: [boost] [cmake] Minimum viable cmakeification for Boost
From: paul (pfultz2_at_[hidden])
Date: 2017-06-21 15:47:27

On Wed, 2017-06-21 at 05:03 +0300, Peter Dimov via Boost wrote:
> Paul Fultz II wrote:
> > 
> > > 
> > > On Jun 20, 2017, at 6:33 PM, Peter Dimov via Boost 
> > > <boost_at_[hidden]> wrote:
> > > 
> > > paul wrote:
> > > 
> > > > 
> > > > https://github.com/steveire/BoostCMake
> > > * Latest commit cd85034Â Â on Aug 29 2013
> > > * cmake_minimum_required(VERSION 2.8.11 FATAL_ERROR)
> > > 
> > > I'd think that even Stephen Kelly can't produce idiomatic CMake 3.5Â 
> > > while using CMake 2.8. Although who knows.
> > This is a lie though because it's using cmake 3 features. It won't build 
> > with cmake 2.8 at all.
> Yes, I looked at the code and it's very similar to what is being suggested 
> today:
> 
> https://github.com/steveire/BoostCMake/blob/master/listsfiles/libs/system/CM
> akeLists.txt
> 
> Header-only libraries don't seem to enumerate dependencies though:
> 
> https://github.com/steveire/BoostCMake/blob/master/listsfiles/libs/smart_ptr
> /CMakeLists.txt

Yes, but I think its evolved since then. At least that is what is seems from
watching Effective Cmake talk at C++Now this year. 

Utlimately, some authors will want to make their cmake standalone, which will
require enumerating the dependencies, but we can't have some cmake scripts
that can be used standalone and others that require being invoked with a root
cmake. We need the build scripts to be consistent, and we need to support both
scenarios.