Re: [boost] [signals] msvc-6.5 failures

From: Doug Gregor (dgregor_at_[hidden])
Date: 2006-12-04 10:10:33

On Dec 4, 2006, at 9:32 AM, Joaquín Mª López Muñoz wrote:

>
>
> Douglas Gregor ha escrito:
>
>> On Dec 2, 2006, at 6:03 AM, Vladimir Prus wrote:
>>
>>>
>>> There are four failures of the signals library on msvc-6.5, like  
>>> this
>>>
>>>       http://tinyurl.com/tc4et
>>>
>>> All are access violations on runtime.
>>
>> This is really weird... I haven't touched the Signals source in
>> months, and this only popped up with the switch to BBv2.
>>
>>> Doug, do you think you'll have the time next week to look into
>>> this? Anybody else
>>> is willing to debug the problem?
>>
>> The trick is to find a copy of MSVC 6 that I can play with :(
>
> I've got a copy myself: I've manually built one of the failing tests
> (deletion_test.cpp) and everything's playing smoothly here... One  
> quirk I've
> noticed, don't know if it can make a difference, is that the  
> associated
> Jamfile.v2 includes a dependency on boost_test_exec_monitor that is  
> not
> really needed.

I'd hope that it wouldn't matter...

> Do you want me to try something in order to ellucidate what's going  
> on?

I don't even know what to try :(

If you test with "bjam" in libs/signals/test, do you get the same  
failures as the regression tester?

        Cheers,
        Doug