Re: [jamboost] Re: [boost] bjam variants, subvariants and default build

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2002-07-16 20:13:14

----- Original Message -----
From: "rogeeff" <rogeeff_at_[hidden]>
To: <jamboost_at_[hidden]>
Sent: Tuesday, July 16, 2002 8:59 PM
Subject: [jamboost] Re: [boost] bjam variants, subvariants and default
build

> --- In jamboost_at_y..., "David Abrahams" <david.abrahams_at_r...> wrote:
> > > > Not implied by lib, nor dll. Only <shared-linkable>true is
> implied
> > > by dll.
> > >
> > > How does it related to <target-type>?
> >
> > <target-type> encodes the type of target being built. It's
> unrelated to
> > <runtime-link>, because in general you can link any kind of target
> to
> > either static or dynamic runtime libraries.
>
> I meant: how does it related to <shared-linkable>?

Well, some target types require that their objects be shared-linkable.
Which ones is left as an exercise for the reader <wink>.

> > Because it wasn't designed to work that way, and probably won't
> work?
>
> Could we make it work?

If by "we" you mean "you", then that's fine with me. Those of us working on
Boost.Build have declared a moratorium on working on new features for v1,
so we can concentrate on v2. Rene has been graciously applying bugfixes,
but even that is very slow going because the code is so tangled -- a big
reason for the v2 rewrite.

> Or you don't think it's resonable request?

Well, I guess the feature request seems a little ill-conceived. I wouldn't
add a feature to change the default variant just to reduce subvariant
grist. I'm not even sure if that would work. Subvariant grist is not
supposed to include any of the properties which are implied by the variant
anyway. In v2 subvariant grist is pretty well minimized.

-Dave