|
|
Boost Testing : |
From: David Abrahams (dave_at_[hidden])
Date: 2007-07-03 19:21:00
on Tue Jul 03 2007, Joaquin M Lopez Munoz <joaquin-AT-tid.es> wrote:
> David Abrahams <dave <at> boost-consulting.com> writes:
>
>>
>>
>> on Tue Jul 03 2007, Joaquin M Lopez Munoz <joaquin-AT-tid.es> wrote:
>>
>> > Not much, really. I've seen it appear and go at random places
>> > in MSVC 8.0 regression testers for more than a year if my memory
>> > serves me well.
>>
>> How do you know? I mean, what characterizes this problem?
>
> The error message:
>
> LINK : fatal error LNK1000: unknown error; consult documentation
> for technical support options
>
> I've grown used to spot it, in my lib and others, because it
> appears suddenly in otherwise healthy libs, so we've met several
> times during my cursory looks for regressions to fix in this last
> year of code freeze.
>
>
> [...]
>> Well it sounds from the above like you're saying that "the problem"
>> is always an ICE.
>
> An ILE, if you allow me the pedantry.
Okay. Well someone at MS had this to say:
------
It is possible the LNK1000 errors are already being reported to
Microsoft.
To be certain, you can use the linker option:
/ERRORREPORT:{NONE|PROMPT|QUEUE|SEND}
Here is the documentation for the option:
http://msdn2.microsoft.com/en-us/library/ms235602(VS.80).aspx
And this link has more information on this option relevant to both the
compiler and linker:
http://msdn2.microsoft.com/en-us/library/ms173502(VS.80).aspx
------
Maybe we should enable that option for regression testing with VC++.
-- Dave Abrahams Boost Consulting http://www.boost-consulting.com The Astoria Seminar ==> http://www.astoriaseminar.com