Re: [Boost-build] Sunpro tests and toolset question

From: David Abrahams (dave_at_[hidden])
Date: 2006-07-12 19:19:52

Vladimir Prus <ghost_at_[hidden]> writes:

> On Wednesday 12 July 2006 13:38, John Maddock wrote:
>> There are quite a few TR1 tests that are failing with sunpro, that appear
>> to be an include path problem.  I believe this was working OK with bbv1,
>> but the include paths appear to be coming out in the wrong order with the
>> current bbv2.
>>
>> A typical list of errors here: http://tinyurl.com/pcwf5 suggests that the
>> "real" version of <memory> is getting included rather than the Tr1 version.
>> The command line shows that the TR1 headers are in the path, but *after*
>> the Boost include path, and while I'm not completely sure, I believe this
>> is the problem (I've seen issues like this before, but they're really hard
>> to diagnose without spending way too much time studying pre-processor
>> outputs). So... anyone any ideas why the toolset is injecting the boost
>> path ahead of other <include> directives?
>
> Hi John,
> as far as Boost.Build V2 is concerned, boost include directory and TR1 include 
> directory as the same, they are both specified using <include> feature, and 
> no specific order is guaranteed.
>
> I think V1 has <sysinclude>, but I never seen any explanation how it's 
> different from <include>. Probably <sysinclude> appear before <include> and 
> that's making it working in V1.

<sysinclude> is designed to search angle-bracket includes first.  Some
compilers make a distinction between "" and <>.  CodeWarrior is one of
them.  If Sun does the same I believe this explains all of the
Boost.Python failures with SunPro.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com