From: David Abrahams (dave_at_[hidden])
Date: 2004-07-02 10:15:38
Vladimir Prus <ghost_at_[hidden]> writes:
> David Abrahams wrote:
>> > This might work for <os>, but not the <toolset> -- since we really have
>> > to have only a single value from a predefined set for it. The
>> > "use-property" rule will have no problems with it.
>> An attribute still sounds like a good idea. You don't have to make
>> it allowable to have multiple unknown values outside of conditions.
> I'm afraid I don't understand. Do you mean that with attribute which can
> specify that values of specific feature need to be checked when used inside
Well, I'd make it an attribute that says certain features *don't*
need to be checked, but that's the basic idea.
> I can see that this attribute can be usefull, but on the other
> hand we currently only have <os> and <toolset> for which such handling was
> requested, and those features might well be handled by 'use-property' --
> which we seem to need anyway.
I'm not convinced we need it, if we have the right attributes.
> I'm just trying to minimize the amount of new things to add ;-)
Bravo, but let's also minimize the amount of things users have to
-- Dave Abrahams Boost Consulting http://www.boost-consulting.com
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