[Uml-devel] Fwd: Re: umbrello compilation errors in RC1]

[Jean Vittor]

I haven't look the incriminated code but I guess that defining
STATIC_ASSERTION_FAILURE<false> will probably not solve the problem.

Please take a look at this page (
http://www.boost.org/doc/libs/1_37_0/doc/html/boost_staticassert.html )
which explains the usage of |BOOST_STATIC_ASSERT|.

STATIC_ASSERTION_FAILURE<false> is expected not to be defined in order
to generate error messages at compilation time.


Jean

Jonathan Riddell a écrit :
> ----- Forwarded message from Bernd Brandstetter <bbrand at freenet.de> -----
>
> From: Bernd Brandstetter <bbrand at freenet.de>
> To: kde-devel at kde.org
> Subject: Re: umbrello compilation errors in RC1
> Date: Thu, 22 Jan 2009 11:27:41 +0100
> Reply-To: kde-devel at kde.org
>
> On Tuesday 20 January 2009, Jonathan Riddell wrote:
>   
>> On Mon, Jan 19, 2009 at 06:08:31PM -0500, Orville Bennett wrote:
>>     
>>> I got this while compiling umbrello under kdesdk for the mac. the kuml
>>> list doesn't look like it's been active since 2002-ish so I'm hoping
>>> someone here can forward this to the pertinent people. Sorry for the
>>> noise. CC Me please.
>>>       
>> What version of boost are you using?  This is probably
>> http://bugs.kde.org/177755
>>
>> Try the svn version and let us know if that helps.
>>
>> Jonathan
>>     
>
> Maybe the patch associated with the above bug entry solves the compilation 
> problem but changing a comparison from >= to == is a bit suspect to me. 
> Will this really work in all cases?
>
> After all, gcc complains about 'boost::STATIC_ASSERTION_FAILURE<false>' not 
> being defined. The file preprocessorlexer.h contains the following section 
> near the beginning:
>
> #ifdef Q_CC_MSVC
> // this template isn't available in boost/static_assert.hpp:50
> // but msvc needs it (boost 1.37.0)
> namespace boost {
>   template <> struct STATIC_ASSERTION_FAILURE<false> { enum { value = 
> 0 }; };
> }
> #endif
>
> Simply removing the #ifdef Q_CC_MSVC in order to always explicitly define 
> the missing type resolved the problem for me.
>
> Regards,
> Bernd
>  
>   
>>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
>>>