|
|
Boost Testing : |
From: David Abrahams (dave_at_[hidden])
Date: 2005-06-16 05:47:37
"Victor A. Wagner Jr." <vawjr_at_[hidden]> writes:
> At 14:54 2005-06-15, you wrote:
>>"Victor A. Wagner Jr." <vawjr_at_[hidden]> writes:
>>
>>Speaking of which, how are we going to resolve that change you made
>>recently to the msvc toolset that enables asynchronous EH everywhere?
>
> I'm open to suggestions.
> upside:
> errors_handling_test now "passes" for vc-8_0
> downside:
> incerease in size and running time (presumably)
I'm not sure how many people are using BBv1 as a development tool, so
there's a chance it might be okay. That said, I'm reluctant to make
the change without knowing for sure.
> since I've always felt it's the programmer's job to "make it right, THEN
> make it small/fast/whatever" altering the exception handling model in the
> compiler seemed the only choice.....
That's the right choice; it just should have been done on a
case-by-case basis.
> afterall, IMO, divide by 0 _should_
> be catchable in a C++ program (perhaps an extension to the core language is
> in order?).
That's not an argument I'm ready to have so early in the morning :)
> so, what to do? I suppose finding a solution in the test-framework,
> build-system, or the bjam files would be preferable, but bbv1 is
> hopefully going away, I'll learn bb when v2 is "official".
>
> I haven't heard any screams coming from the testing community since I've
> made the change to the toolset and since it fixed a problem that nobody
> seemed to want to look at ...... <shrug>
>
> someone else's call..... I'm only a "tester"
Was it only that one test that was failing? We could easily add an
<asynch-exceptions> feature and set it to "on" in that test's
requirements.
-- Dave Abrahams Boost Consulting www.boost-consulting.com