Re: [boost] [Report] 2992 failures on trunk (2007-09-26)

From: Doug Gregor (dgregor_at_[hidden])
Date: 2007-09-27 10:30:48

On Sep 27, 2007, at 10:09 AM, Rene Rivera wrote:

> Markus Schöpflin wrote:
>> Douglas Gregor wrote:
>>
>>> Nikolay Mladenov wrote:
>>>> This test is not using the most recent trunk files.
>>>>
>>>>
>>>>> |python|
>>>>>   pytype_function:  hp_cxx-71_006_tru64
>>>>>
>>> The results on that whole platform are way out of date. I've removed
>>> them from the reports.
>>
>> Umm, they're only a bit more than a week old. Is that considered  
>> way out of
>> date? If this is the case, I can as well give up regression  
>> testing for
>> Tru64, as it takes nearly seven days for both compiler versions to  
>> complete
>> the full regression cycle.
>
> I think Doug was a bit premature in removing those tests.

Ouch, sorry. I didn't realize it was normal for the Tru64 tests to  
run for 7 days.

In that case, we cannot consider Tru64 a release-critical platform,  
and it should be removed from the release view of the regression  
tests. Any platform with 7-day cycle times will lengthen a release  
schedule. Of course, regression testing for the platform is still  
important, and we'll still try to get everything we can working for  
that compiler.

        - Doug