Re: [Boost-testing] Useless failures reports/broken testers

Subject: Re: [Boost-testing] Useless failures reports/broken testers
From: David Deakins (ddeakins_at_[hidden])
Date: 2008-12-23 23:21:28

Gennadiy Rozental wrote:
> I've checked in some iffy code into Boost.Test and impatiently waiting 
> for regression results to appear here:

> 
> * VeecoFTC (msvc-8.0~wm5~stlport5.1)
> 
> Reports failure in library build, but link to the library build output 
> shows success
> 

I'm not sure if there was a hiccup in the report status previously, but 
currently the VeecoFTC Windows Mobile results are correctly showing that 
Boost.Test fails to build because putenv (used in putenv_impl in 
boost/test/utils/runtime/config.hpp) does not exist on this platform. 
Previously, runtime/config.hpp was not included as commonly in the 
regression tests, but the recent changes to 
boost/test/impl/unit_test_parameters.ipp now cause 
boost/test/utils/runtime/env/variable.hpp to be included.  This 
eventually draws in runtime/config.hpp and triggers the compiler error. 
  Attached are proposed patches to

boost/test/utils/runtime/config.hpp,
boost/test/utils/runtime/env/environment.ipp, and
boost/test/utils/runtime/env/fwd.hpp

that I believe should resolve the WinCE problems.  Let me know if you 
see any issues with them.

-Dave