From: Vladimir Prus (ghost_at_[hidden])
Date: 2005-09-16 08:15:51
On Wednesday 14 September 2005 23:52, Alexey Pakhunov wrote:
> Vladimir Prus wrote:
> > What I don't understand in your patch is the "registered" state. Is that
> > a state where toolset is known to Boost.Build internals but not usable by
> > the user? Or is it state where toolset is already usable by the user, but
> > maybe be still reconfigured?
> This is the state when a toolset configuration is detected (by
> autodetection code) but 'toolset.using' for this configuration wasn't
> yet called.
1. How a toolset can get in this state? After
configure gcc : all ;
2. What can I do with toolset in that state? Can I request such toolset on
command line? Say:
3. I'd expect toolset.using to be called for all invocations of "configure".
Do you mean "configure with explicit version"?
> > In class docs, your talk about "condition". Here you use "id" without
> > saying what it is.
> I removed the word 'condition'. It can be any kind of unique ID. Toolset
> conditions one of possible options.
> Other comments were added as well. The updated patch is attached.
> > Everything else looks fine, but can be roughly outline the further
> > direction?
> Andrey posted quite detailed roadmap. :-) So I can add only that the
> next changes will be:
> > 5. new configure rule
> > 6. "configure all" syntax
> > 7. "version registration" framework
Can you send them all now? I'll still review and commit them one-by-one but ti
will help me to see the next step.
-- Vladimir Prus http://vladimir_prus.blogspot.com Boost.Build V2: http://boost.org/boost-build2
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