[boost] Re: RC_1_31_0 regressions in multi_array, counting_iteratorandlexical_cast

From: David Abrahams (dave_at_[hidden])
Date: 2004-01-20 08:29:28

Martin Wille <mw8329_at_[hidden]> writes:

> David Abrahams wrote:
>> Martin Wille writes:
>> 
>>>Hi,
>>>
>>>We have currently have a few regressions on Linux:
>>>
>>>
>>>conversion: lexical_cast_test fails for intel 7.1 and intel 8.0
>>>
>>>multi_array: several tests fail for gcc-2.95.3-stlport-4.5.3
>>>
>>>utility: counting_iterator_example fails for gcc-2.95.3-stlport-4.5.3
>> Don't know what to say about these last two.  "(remainder deleted
>> because of excessive size)" makes it kinda hard to even guess at the
>> multi array problem, and counting_iterator_example?
>
> I don't find this text in the current logs.
>
>
>>    /home/BOOST/head-regression/boost/boost/iterator/indirect_iterator.hpp:127: `x' undeclared (first use this function)
>> Gimme a break!
>>   template <class Iter>
>>   inline
>>   indirect_iterator<Iter> make_indirect_iterator(Iter x)
>>   { //------------------------------------------------^ !!!
>>     return indirect_iterator<Iter>(x);
>>   }
>> So, I don't know what to say about that.  My impression is that
>> gcc-2.9x is just choking on the long symbol names we get with STLPort
>> debug mode.
>
> I agree.
>
>>              You might try turning that off.
>
> Wouldn't the debug mode be superior for the tests the
> compiler doesn't choke upon?

Yes, but it's fair to turn it off for a single test.

> I think we should just add an explicit failure markup
> for that test. This obviously is a compiler problem.

Either way is OK for me.

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com