Boost logo

Boost-Build :

Subject: Re: [Boost-build] direct_request_test.py
From: Vladimir Prus (ghost_at_[hidden])
Date: 2011-12-09 10:40:23


On 14/12/11 00:47, Steven Watanabe wrote:
> AMDG
>
> On 12/09/2011 04:05 AM, Vladimir Prus wrote:
>> On 09/12/11 08:19, Steven Watanabe wrote:
>>> AMDG
>>>
>>> This test is not run by the test suite and
>>> currently fails. The reason is that it
>>> is testing that for
>>>
>>> Jamfile.jam:
>>>
>>> exe test : test.cpp :<variant>debug ;
>>>
>>> b2 release define=MACROS
>>>
>>> the property<define>MACROS is /not/ used
>>> to build the target test, because release
>>> conflicts with<variant>debug.
>>>
>>> The current behavior is to apply all the
>>> command line features independently.
>>> I'm presuming that this is the correct
>>> behavior, but I wanted to check before
>>> I update the test.
>>
>> I wanted define=MACROS to be used in this case. The idea is that free
>> features on the command line
>> have to apply to all targets built, since they are not propagated and
>> would not have any
>> effect otherwise. Does this make sense?
>>
>> targets.jam has this to say:
>>
>> # Apply free features from the command line. If user said
>> # define=FOO
>> # he most likely wants this define to be set for all compiles.
>> property-set = [ $(property-set).refine
>> [ build-system.command-line-free-features ] ] ;
>>
>> no idea why this is not working -- it surely worked when I added the
>> test although it's
>> depressing that I forgot to include it in test_all.py
>>
>
> I guess I wasn't clear. It /is/ working.
> It's the test that's wrong.

Not sure I get it. define=MACROS should be used to build test.cpp even in this setup, for all
I see.

Thanks,
Volodya


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