Re: [Boost-testing] [boost] [Report] 0 regressions on RC_1_34_0 (2007-04-13)

From: Martin Wille (mw8329_at_[hidden])
Date: 2007-04-23 13:00:36

Roland Schwarz wrote:
> Thomas Witt wrote:
>> Roland,
>>
>> please post questions like this to the testing list.
>>
>> Thomas
>>
>> Roland Schwarz wrote:
>>> See e.g.
>>> http://tinyurl.com/2689bg
>>>
>>> To me this looks as if the report generator is still hiding some errors.
>> IIUC this is marked fail expected.
> 
> I still feel misunderstood, and I am getting somewhat tired, because I 
> already explained this somewhere else. (Search the testing list for
> 'False positives?')
> 
> On the other hand if none else think this is an error, it possibly is
> just my inabilty to understand why a test that is passing has an exit
> code != 0.
> 
> In this case I would be glad if someone can explain.

Roland, I think you're rightfully touching a sore point here.

An exit status of 1536 suggests an error, even more so under the light
of the preceding assertion failure. This test should get reported as
(expectedly) failing, not at all as passing.

This should get fixed before we release, because it is unclear whether
we're currently missing other failures.

Are you running tests incrementally? If so, could you then please remove
the directory
X:\boost-regression\gcc-d\results\boost\bin.v2\libs\regex\test\regex_regress_threaded.test\gcc-cygwin-3.4.4\debug\python-2.4\target-os-cygwin\threading-multi\
before the next run?

Regards,
m