Re: type_traits: fixed is_POD, testing methodology questions

From: rogeeff (rogeeff_at_[hidden])
Date: 2002-02-21 20:47:32

--- In boost_at_y..., "David Abrahams" <david.abrahams_at_r...> wrote:

> About the testing methodology: It seems as though the tests are 
engineered
> so we won't know when something gets fixed. Tests which are 
expected to fail
> to compile on a given compiler never actually happen. The number of 
failures
> reportedly expected for a compiler (e.g. 
msvc/object_type_traits_test) may
> far exceed the number set in the expected_failures variable for any 
given
> test. What can we do so that the status of a given trait on a given 
compiler
> is more-reliably reported? We have everything we need in 
status/Jamfile in
> order to get fine-grained and accurate reports.
> 
> I also made the following change. It causes tests to fail if they 
don't
> match their expected number of failures exactly. I needed to do 
that in
> order to see that the change fixed anything:
new Boost Test Library is also checking number of failures expected 
using less. Would you prefer equal?

Gennadiy.

[...]
> +---------------------------------------------------------------+
>                   David Abrahams
>       C++ Booster (http://www.boost.org)               O__  ==
>       Pythonista (http://www.python.org)              c/ /'_ ==
>   resume: http://users.rcn.com/abrahams/resume.html  (*) \(*) ==
>           email: david.abrahams_at_r...
> +---------------------------------------------------------------+