Re: [jamboost] Alternatives and underfined feature values

From: Zbynek Winkler (zw-bjam_at_[hidden])
Date: 2005-02-03 17:31:37

Vladimir Prus wrote:

>I'm looking at http://zigzag.cs.msu.su:7814/scarab/issues/id/BB83
>which says that if you write:
>
> alias a : a_gcc.cpp : <toolset>gcc ;
> alias a : a_msvc.cpp : <toolset>msvc ;
>
>and don't initialise both gcc and msvc toolset, you'll get errors. This fact 
>makes target alternatives not so usefull.
>
>I see three approaches:
>
>1. Do not validate features here at all.
>2. Introduce special syntax for properties used for alternative selection:
>
> alias a : a_gcc.cpp : when <toolset>gcc ;
> alias a : a_msvc.cpp : when <toolset>msvc ;
>
>and don't validate features specified in "when" clause. This is related to
>http://zigzag.cs.msu.su:7814/scarab/issues/id/BB78, which says we need 
>explicit control over what features are used for alternative selection.
>3. Allow the user to say that <toolset>msvc is valid, without initializing the 
>toolset.
>
>Any opinions on the best approach.
> 
>
I like #3. It seems to be the cleanest of your proposals. In fact I 
think that all toolsets shiped with BB2 should be valid by default. The 
same goes for the <os> feature that has the same problem.

Zbynek

-- 
http://zw.matfyz.cz/ http://robotika.cz/
Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic