Re: [Boost-testing] expected failures test_thread and test_barrier_lib

From: David Abrahams (dave_at_[hidden])
Date: 2006-12-15 08:38:16

Aleksey Gurtovoy <agurtovoy_at_[hidden]> writes:

> Roland Schwarz writes: 
>> The test_thread for msvc-7.1 and test_barrier_lib for msvc-8.0 are
>> marked as expected to fail.
>> 
>> Is this still the case?
>> 
>> Which errors are expected?
>> 
>> Can we remove the expectation?
>
> If you look at 1.32 results (which is when the markup was added) at
> http://engineering.meta-comm.com/boost-regression/1_32_0/developer/thread_release.html
> you'll see that "test_thread" was passing for everyone but us
> ("metacomm"), hence the note:
>  
>    [ Aleksey Gurtovoy ] The failure is probably caused by the test
>    code, harness, or configuration. Thus, it may not affect users of
>    the library.

I think that note leaves something to be desired.  Any reader who
considers using the library in that context would like to know why you
think it's "probably" true.  Couldn't you add that it passed for all
but one tester, and the reasons for the one failure couldn't be
explained?

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