From: Vladimir Prus (ghost_at_[hidden])
Date: 2002-07-15 08:49:40
David Abrahams wrote:
> > It appears that I prefer to disallow cases not explicitly needed, so that
> > can enable them later, while you have exactly opposite approach. Don't
> > which is best, althought yours seems easier to implement.
> In general, I agree with you especially when the language is as liberal as
> Jam's. However, features won't be defined often and it's OK to ask people
> to exercise some care in this case. A bad experience with building in the
> wrong arbitrary limitations is leading me to be a little more liberal, and
> a desire to keep the implementation straightforward (so we can debug it)
> helps too.
Okay. Agreed to choose this way now.
> > > I just don't think we're ready to design attribute groups yet. Partly
> > > that's because your terms "principal", "minor", and especially "trait"
> > > don't resonate for me. Maybe if you listed all our existing (v1)
> > > and a few anticipated v2 features, and tried to categorize them this
> way it
> > > would become clearer to me...
> > Can we try splitting "principal -> ... " axis into
> > "subvariant-dir" and "main-target-dir" attributes and proceed with this
> > arrangement?
> Sure, that sounds fine to me. Just to keep people from having to type
> subvariant-dir all over the place, though, I'd rather that was the default,
> so we have
> Oh, wait: they're orthogonal! Imagine property <foo>bar is in the build of
> target baz:
> no-subvariant:0 1
> main-target-dir \------------------------------------------
> 0 | .../bin/foo-bar/ .../bin/
> 1 | .../bin/baz/foo-bar/ .../bin/baz/
> I see no problem here.
Well, since "main-target-dir" means "create a dir corresponding to main
target when properties of target differ project requirements",
main-target-dir and subvariant dir are just two methods to make sure we don't
fix targets with different properties together. Therefore, I don't understand
the meaning io "no-subvariant = 0, main-target-dir = 1". Why put name target
name in target path when we already have subvariant dir there?
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk