|
Boost-Build : |
From: Pedro Ferreira (pedro.ferreira_at_[hidden])
Date: 2004-12-14 06:13:59
Em 14 Dec 2004, às 07:57, Vladimir Prus escreveu:
>
> On Monday 13 December 2004 23:33, Toon Knapen wrote:
>> Pedro Ferreira wrote:
>>> My vision is:
>>> - create a 100% Python Boost.Build library on top of SCons
>>> - create a parser for the bjam language
>>
>> Just wondering: do we really need backward compatibility with the bjam
>> language? Don't get me wrong, I've already invested lots of time in
>> getting my Jamfile's right so I have a vested interested in
>> compatibility.
>>
>> But OTOH I think that long-term it might not be the best option
>> because
>> we need to have a parser for bjam. Why not use python language (like
>> scons) directly in the build-scripts ? The strong point of boost-build
>> are the concepts in the user-interface. And these are what we need to
>> 'port' on top of scons. But I would not insist to write
>>
>> exe foo : foo.cpp : <include>../include ;
>>
>> instead of
>>
>> create_exe( foo, 'foo.cpp', '<include>../include' )
>
> I think we already have parser for Jamfile -- the one in Jam.
> Integrating that
> with Python solution (whatever specific design is used) should be
> simple.
>
> OTOH, I believe that preserving Jamfiles is an important goal. It's
> kind of
> "marketing" point. Even if Boost.Build is not yet 2.0-final and we
> made no
> promises, breaking compatibility will cause much troubles for users.
> Even
> imagine that a project works fine with 2.0-m11 but does not work with
> (say)
> 2.0-m12. Everybody will be upset with how Boost.Build is incompatible
> with
> itself.
Agreed. I think that whatever way we go, compatibility must be
maintained.
Pedro
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