On Wed, 14 Sep 2011 17:32:37 -0700
Brian Harring <ferringb@...> wrote:

> On Thu, Sep 15, 2011 at 12:15:59AM +0100, Ciaran McCreesh wrote:
> > On Wed, 14 Sep 2011 11:19:35 -0400
> > Mike Frysinger <vapier@g.o> wrote:
> > > however, why wont this work sanely in src_* or pkg_* funcs ?  the
> > > env there is the one constructed by the PM which includes the
> > > merged IUSE values.
> > 
> > It's not. We deliberately don't specify that the PM passes fixed up
> > values for IUSE, DEPEND etc back into the ebuild, since Portage's
> > behaviour for global variables has varied considerably over the
> > years.
> 
> That's not varied; the implementation for eclass/ebuild stacking has 
> been pretty consistant for phase funcs.
> 
> Either way, candidate for eapi5 if it's worth ensuring it's
> accessible to the phases...

Is that actually useful? Isn't that just a waste of time serializing
all that data back, of which most isn't suitable for non-specialized
parsing?

If we really need to access ${IUSE}, I think we should go for something
IUSE-specific.

-- 
Best regards,
Michał Górny