Re: [Boost-testing] [accumulators] spurious failures?

From: K. Noel Belcourt (kbelco_at_[hidden])
Date: 2008-01-08 17:34:16

On Jan 8, 2008, at 3:27 PM, John Pavel wrote:

>
>> Ah-ha! Yes, iterator debugging makes some of the accumulator tests  
>> run
>> virtually forever. I'll fix this in the jamfile. Thanks. We'll see if
>> that makes the spurious failures go away.
>
> Well I managed to run all the tests in stand-alone mode and they  
> worked OK,
> but some did seem to run over the 5m limit that I think is the  
> default.  The
> regression testing system should log a timeout if the test  
> overruns, rather
> than leaving us to guess.

I believe bjam does kill processes that run past the time limit.   
What was the time limit and how far past the time limit did the test  
run before terminating?  Also, a diagnostic is issued if any process  
killed for exceeding the time limit.  Was no diagnostic issued?

-- Noel