Re: [Boost-testing] [repost][test][msvc-6.5_stlport4][speedsnail-msvc-dp] should we force static linking for Boost.Test on MSVC++ 6.5 + STLport?

From: Vladimir Prus (ghost_at_[hidden])
Date: 2007-02-15 15:32:17

"JOAQUIN LOPEZ MU?Z" wrote:

> Any feedback on this, please? Thank you!
> 
> Last Monday I wrote:
> 
> Hello,
> 
> A couple of weeks ago we decided about the convenience of
> forcing static linking of Boost.Test on MSVC++ 6.5, given
> that DLL linking is definitely not supported for this lib
> on this environment, so that some tests for this lib would
> pass instead of showing as misleading failures. As a result,
> Roland added the following conditional requirement to the
> RC_1_34_0 branch of libs/test/build/Jamfile.v2:
> 
> 
> lib boost_unit_test_framework :
>  $(UTF_SOURCES).cpp : <toolset>msvc-6.5:<link>static ;
> 
> 
> Now, the very same problem is showing for MSVC++ 6.5 +
> STLport in speedsnail-msvc-dp, see for instance
> http://tinyurl.com/3yub7k .
> 
> 
> Shouldn't we then extend the <link>static requirement to
> also cover this case? I don't know whether the following
> would do (are wildcards supported in this context?):
> 
> 
>   <toolset>msvc-6.5*:<link>static
> 
> 
> or else we need an extra requirement or a smarter
> construct.

The above won't work, so you either need a separate requirements
for msvc-6.5 with stlport4 or employ "conditional requirements"
mechanism. Using a second requirement for stlport might be easier
and more direct solution.

Are you in position to try it yourself or want Roland and me to
do this?

- Volodya