[PATCH][libstdc++-v3 parallel mode] Ambiguities (incl. std::mismatch)

[François Dumont]

Thanks for pointing this.

I miss this thread because I am working on a coming branch modifying the 
way debug algos are implemented. In this version I have now 3 versions 
of equal, the normal one, the debug one and the parallel one. So I will 
be very careful with those qualification issues. I will make a dedicated 
mail for that soon.

Regards


On 01/25/2011 08:22 PM, Paolo Carlini wrote:
> Hi,
>    
>> I have come to the conclusion that my patch from 2010-03-22 was mostly
>> bogus (r157643), and deactivated parallel execution for the affected
>> routines in certain cases.  The namespace _GLIBCXX_STD_P always
>> suggests "parallel", but it is the very opposite :-(
>>
>> I have added test cases (for conformance) to existing files since this
>> is the problem of mixing implicit and explicit calls to the parallel
>> mode again.
>>
>> Tested x86_64-unknown-linux-gnu: No regressions
>>
>> Please approve for mainline and gcc-4_5-branch, also waiting for the
>> feedback of François.
>>      
> Thanks for sorting out this issue. Patch is Ok for mainline now and in a
> week or so for 4_5-branch too, assuming of course you tested both make
> check and make check-parallel.
>
> A nit: testcases too should be wrapped to 80 columns.
>
> Thanks again,
> Paolo.
>
>