Re: [Boost-build] Generating multiple CPP/H targets from a singlesource

Subject: Re: [Boost-build] Generating multiple CPP/H targets from a singlesource
From: Fabien Chêne (fabien.chene_at_[hidden])
Date: 2010-08-17 10:24:46

2010/8/17 Juraj Ivančić <juraj.ivancic_at_[hidden]>:
> On 17.8.2010 14:57, Fabien Chêne wrote:
>
>>> You can add this to the user-config.jam:
>>>
>>> modules.poke : ARGV : [ modules.peek : ARGV ] --abbreviate-paths ;
>>
>> I've added the above line in my user-config.jam. It updates ARGV but
>> unfortunately, it does not affect dir names.
>
> Are you sure? I have been using this trick for ages and never had any
> problems with it. How can you tell that it updates ARGV?

echo [ modules.peek : ARGV ] ;
modules.poke : ARGV : [ modules.peek : ARGV ] --hash ;
echo [ modules.peek : ARGV ] ;

Here is the bjam output when adding the below lines in user-config.jam:
/xusers/fc01260/Bjam/boost-build/engine/src/bin.linuxx86_64/bjam
/xusers/fc01260/Bjam/boost-build/engine/src/bin.linuxx86_64/bjam --hash

So, it does something for sure, but it does not seem to be correctly
forwarded to Boost.Build.

[...]
>> About compile timings, it behaves the same than the normal build, even
>> if both options should save memory.
>
> Not sure what you mean by that. These two options only change output
> path names, nothing else.

As paths are slightly reduced, I was hopping for a small memory
saving. I believe bjam is known to be slow because of its memory
usage, which mainly involves huge strings. Ok then I'll wait for the
python port :-)

-- 
Fabien