Re: [boost] Boost regression notification (2005-06-26)

From: David Abrahams (dave_at_[hidden])
Date: 2005-06-26 18:33:14

Rene Rivera <grafik.list_at_[hidden]> writes:

> Douglas Gregor wrote:
>> Boost regression test failures
>> ------------------------------
>> Report time: 2005-06-26T15:40:21Z
> [...]
>> 717 failures in 22 libraries:
>
> Don't know what happened, but there are a few new CW-8.3 regressions now...
>
>> |numeric/conversion|
>>   converter_test:  cw-8_3
>>   udt_support_test:  cw-8_3
>
> Pass for me locally.
>
>> |preprocessor|
>>   seq:  cw-8_3
>
> Not easily testable on it's own, as there's no test/Jamfile for BPP.
>
>> |type_traits|
>>   decay_test:  cw-8_3
>
> This change:
>
> <http://cvs.sourceforge.net/viewcvs.py/boost/boost/libs/type_traits/test/decay_test.cpp?r1=1.1&r2=1.2&sortby=date>
> <http://tinyurl.com/ap7uc>
>
> Seems to have caused it. But Dave's comment on the change doesn't say 
> which "Broken compiler" the change was for.

Sorry, it was one of the VC++s; 6 I think.  I figured omitting the
void would work anywhere.  Well, you're probably safe by optionally
including the void when on CW, yes?

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