Re: [boost] [spirit.phoenix] Phoenix test coverage

Subject: Re: [boost] [spirit.phoenix] Phoenix test coverage
From: Thomas Heller (thom.heller_at_[hidden])
Date: 2011-02-07 17:02:02

Jim Bell wrote:

> On 1:59 PM, Hartmut Kaiser wrote:
>>> >From names in the spirit regression test matrix
>>> <http://www.boost.org/development/tests/release/developer/spirit.html>,
>>> I see these four with phoenix in the name: dedent_handling_phoenix,
>>> lex_dedent_handling_phoenix, lex_set_token_value_phoenix,
>>> set_token_value_phoenix.
>>>
>>> When I search for phoenix in all libs/spirit/test code, I see references
>>> in 104 files, so tests span more than just above four.
>>>
>>> But how well do phoenix's tests span its content?  Is this documented
>>> somewhere?
>> Tests for Phoenix V2 are not run at all, IIUC. This will change only
>> when/if Phoenix V3 has been moved to trunk after its mini-review. The
>> tests you're referring to are Spirit specific tests using Phoenix
>> features.
> 
> Looking at what's already released, libs/spirit/phoenix/test, I see what
> looks like a full suite of tests, and when I run bjam on them I get 45
> tests passed (on MinGW gcc-4.4/win-32) and one failure.
> 
> So the full compliment of tests is there and likely known to work on a
> couple platforms, just not run by the regression mechanism.
> 
> Thomas and Vincente -- thanks for the v3 input, too.
> 
> For the record, here's the failure (MinGW gcc-4.4/win-32):
> ====== BEGIN OUTPUT ======
> Direct Init -1, 3.14
> Abracadabra 987, 9.87e+032
> scope\dynamic_tests.cpp(68): test 'clos.real() == clos.num() * 1e30'
> failed in function 'int main()'
> Hello World 579, 10.2478
> scope\dynamic_tests.cpp(74): test 'clos.real() == clos.num() / 56.5'
> failed in function 'int main()'
> Assertion failed: called_report_errors_function, file
> ..\..\..\../boost/detail/lightweight_test.hpp, line 42
> 
> This application has requested the Runtime to terminate it in an unusual
> way.
> Please contact the application's support team for more information.
>  
> EXIT STATUS: 3
> ====== END OUTPUT ======

We know about this bug and it is quite embarrassing ... It is actually a bug 
in the testcase itself :)
I will take care of it ASAP.