Re: [jamboost] Boost.Build V1, build subvariant fix!

From: Rene Rivera (grafik666_at_[hidden])
Date: 2002-04-03 20:05:50

On 2002-04-03 at 07:52 PM, david.abrahams_at_[hidden] (David Abrahams) wrote:

>
>----- Original Message -----
>From: "Rene Rivera" <grafik666_at_[hidden]>
>
>> The problem with subvariant checks and generation is fixed :-)
>
>Wonderful (especially after the 1.55 tweak to boost-base.jam <0.02
>wink>) !
>
>> Specifying a multiple BUILD requests, both single valued and multi
>valued
>> correctly generates all the subvariants, without complaints about
>> incompatabilities.
>>
>> As an example/proof/test of the workings of the fix I ended up adding
>a new
>> feature which has come up again and again... <architecture>. This is
>mainly
>> for test purpouses as I only enumerated the x86, Sparc, and Power
>variants.
>> And only implemented the flag changes in the gcc tool.
>>
>> For fun try specifying "-sBUILD=<architecture>i386/i486i586/i686/k6"
>and watch
>> it generate all those targets :-]
>
>Did you by any chance try mixing orthogonal features? That was one of
>the complaints:
>
>-sBUILD="<architecture>i386/i486 <runtime-link>static/dynamic"

Yes ;-) That's why I added another totally different feature from what was
there already. The above, should, correctly build all 4 variants.

And Jerry did point out another bug I had, which is already fixed :-)

-- grafik - Don't Assume Anything
-- rrivera_at_[hidden] - grafik_at_[hidden]
-- 102708583_at_icq - Grafik666_at_AIM - Grafik_at_[hidden]