[Boost-testing] Re: Compiler status problems with run-fail tests

From: Beman Dawes (bdawes_at_[hidden])
Date: 2005-06-01 09:24:20

"Beman Dawes" <bdawes_at_[hidden]> wrote in message 
news:d7gi4n$vm1$1_at_sea.gmane.org...
> It looks like compiler_status.cpp isn't reporting the results from 
> run-fail tests correctly. That is affecting Boost.Test results, and 
> possibly some other libraries.
>
> I'll try to fix the problem, and also add test cases so that 
> process_jam_log and compiler_status are tested for all combinations of 
> compile/link/run expected to succeed/fail.
>
> It will take a couple of days.

Fixes to compile_status.cpp have be committed to CVS.

New tests have been added to tools/regression/test to cover more cases. Test 
cases for warnings have been expanded to work as desired (i.e.. generate 
warnings) for more compilers.

I'll try to keep an eye on regression test results to make sure there were 
no unintended consequences of the changes. If anyone else spots problems, 
please post something right away. This late in the release cycle we don't 
want regression testing to break.

--Beman