Re: Tried Boost.build v2-m9.1 and loving it

From: David Abrahams (dave_at_[hidden])
Date: 2004-08-22 08:05:11

The following message is a courtesy copy of an article
that has been posted to comp.lang.c++.moderated as well.

Ruby Script <nospam_at_[hidden]> writes:

> Ruby Script wrote:
>> Tried Boost.build v2.0-milestone 9.1 and really loving it.
>> 
>> It's great for one-liner build scripts as well as very complicated 
>> scenarios involving multiple compilers & compiler versions.
>> 
>> The only major problem is the less-than-satisfactory documentation.
>> 
>> Truly amazing build system....once v2 gets decent documentation and a 
>> book or two, it will do to make what subversion will do to cvs.
>
> Well, the initial honeymoon is over.
>
> Trying to find out simple things like printing/examining the CFLAGS or 
> LINKFLAGS value has been too much of a hassle for me--docs doesn't cover 
> this and jamboost newsgroup seems to have many questions piped to /dev/null.

I'm sorry to hear that. You picked a bad time to try Boost.Build v2.
Vladimir Prus, who's been its main developer, is on vacation until the
end of the month. The rest of us are better-versed in BBv1, and have
mostly been waiting for him to get back to answer your questions.
I'll see what I can do, though.

> Guess I'll have to checkout SCons next. I don't know Python (yet) but 
> SCons seems to be better documented and more actively supported by 
> experienced users than bjam v2.

Possibly so. It's a good tool.

Boost as a whole is about to make the transition (after the upcoming
release) to BBv2 and by then I hope to be better versed in the
internals myself. I started the project, but then ran out of time
and, well, the details have gotten away from me.

> Pity, because I have a fairly complex project I can build using 
> boost.build v2 for several compilers & versions simply by passing 
> different command-line args to surprisingly *TINY* bjam scripts. Hoping 
> it doesn't suck up too much time to migrate to SCons (or hundreds of 
> extra lines of build script to do this).
>
> Others willing to dig thru the bjam source files will probably have 
> better luck than I...but if the main reason for switching to a new tool 
> is to save time, why bother?

Heh, that's the problem with prerelease software. BBv2 is
better-designed than BBv1, but it's not released yet so you can't
quite count on the same level of support.

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com