Re: [boost] Re: Regression Tests, Boost.Serialization

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2004-07-23 03:04:46

David Abrahams writes:

> Aleksey Gurtovoy <agurtovoy_at_[hidden]> writes:
>
>> David Abrahams writes:
>>> Martin Wille <mw8329_at_[hidden]> writes:
>>>
>>>>> It would be nice if we could drop serialization on compilers that
>>>>> just aren't
>>>>> going to work.
>>>>
>>>> Right. I once suggested that this should be
>>>> implemented for all libraries. Its nonsense
>>>> to run the tests for libraries which are marked
>>>> as non-working for certain compilers. This should
>>>> be a feature of the build system.
>>>> We don't have that feature, yet. I'm not aware
>>>> of anyone working on it. 
>>>
>>> I'd be willing to try, but I think we might get more bang-for-the-buck
>>> if I set things up so that failed tests don't run until they're
>>> outdated.  
>>
>> That won't help with clean runs, though, and it would be really
>> wonderful to have them speed up a little.
>
> OK.
>
>>> That part can be done entirely within Boost.Build rather
>>> than trying to figure out how to combine some XML markup with
>>> it... 
>>
>> I think duplicating markup in Jamfiles or, preferrably, near them (in
>> some form) won't be too bad. E.g., in the library "test" directory we
>> could have a simple "unusable-tools.jam" which could go like this:
>>
>> unusable-tools = borland-5.5.1  msvc  msvc-stlport ;
>>
>> If we can do something like that, it's even not necessarily a
>> duplication -- for XSL reports, we can always walk through the library
>> directories, collect the markup and transform it into an XML for later
>> processing. In fact, we already to something like this anyway.
>>
>>> unless something in the Jamfile that causes the test to be
>>> skipped for certain toolsets is enough for you.
>>
>> It's the opposite -- a toolset marked as unusable should be
>> skipped. 
>
> What does it mean for the toolset to be marked, 

Sorry, imprecise wording. Of course it's the library that is marked as
unusable with a particular toolset.

> and how does it differ from what I suggested?

You seemed to imply that it's skipping of individual tests for
particular toolsets that is of our interest. I was trying to make a
point that the most important use case is to mark the whole library's
*test suite* as unusable. May be from implementation standpoint there
is no difference -- but since I have no idea, I thought I'd point it
out.

-- 
Aleksey Gurtovoy
MetaCommunications Engineering