Boost logo

Boost :

Subject: Re: [boost] RFC: Separating Boost.Python from Boost
From: Peter Dimov (lists_at_[hidden])
Date: 2015-05-30 09:54:12


Niall Douglas wrote:

> On 30 May 2015 at 11:43, Bjorn Reese wrote:

> > I do not know why anyone want to do that for Boost.Python, but this
> > problem is of a general nature.
> >
> > Every author of a new Boost library has to go through the "birth-pain"
> > of getting their library to build with Boost.Build while it is not part
> > of monolithic Boost yet.
> >
> > Currently, the easiest way is to checkout the Boost codebase and then
> > copy your library in there. A more involved alternative is to cheat
> > Boost.Build into believing that your library is part of the Boost code
> > base by adding lots of symbolic links all around. Neither solution is
> > feasible if you want other projects to use your library.
>
> I'm a bit surprised you say this. AFIO requires exactly one symlink into
> libs/. Modular Boost takes care of all the rest if your Boost library has
> the right directory structure.

Suppose you have Boost in C:\boost-1.58.0, and your proposed Boost.Pumpkin
is in C:\projects\pumpkin.

If you symlink C:\boost-1.58.0\libs\pumpkin to C:\projects\pumpkin, you'll
be able to do this:

    C:\boost-1.58.0\libs\pumpkin\test>b2

and it will run the tests. But what I think Bjorn wants is this:

    C:\projects\pumpkin\test>b2

This presumably would involve a Jamroot either at C:\projects\pumpkin or at
C:\projects that would point to C:\boost-1.58.0.

> On Windows, enabling normal users to create symlinks also gets rid of any
> weird build failure problems. Or you simply copy afio into libs/afio and
> be done with it.

For directories, you can use junctions on Windows, mklink/j.


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