windows build

[Paulo Jorge Guedes]

> -----Original Message-----
> From: Peter Kümmel [mailto:syntheticpp at gmx.net]
> Sent: terça-feira, 27 de Junho de 2006 10:20
> To: kde-buildsystem at kde.org
> Subject: Re: windows build
> 
> Thiago Macieira wrote:
> > Peter Kümmel wrote:
> >> Thiago Macieira wrote:
> >>> It looks like this is being caused by the KDE-side of things. Why
> >>> should kauthorized.obj define the QList<QVariant>::isEmpty() and
> >>> *export* it? That's an inline method defined in qlist.h.
> >> Does it export it? kauthorized has its own instantiation and this one
> >> conflicts with the instantiation exported by qdbus.
> >
> > Isn't this the only reason why it happens?
> >
> > If either symbol were not exported, there would be no conflict.
> 
> Maybe because of a missing inline:
> 
> Index: kconfigbase.h
> ===================================================================
> --- kconfigbase.h       (revision 555266)
> +++ kconfigbase.h       (working copy)
> @@ -1540,7 +1540,7 @@
>  }
> 
>  template <typename T>
> -void KConfigBase::writeEntry( const char* pKey, const QList<T>& value,
> +inline void KConfigBase::writeEntry( const char* pKey, const QList<T>&
> value,
>                                WriteConfigFlags pFlags )
>  {
>  #if KCONFIG_QVARIANT_CHECK
> 
> 
> writeEntry uses a QVariantList which will be exported by kdecore when this
> function is not inline, or I'm wrong.

I tried and I get the same error :(

Paulo

P.S.: kconfigbase.h:1179:

  virtual void setReadOnly(bool _ro) { bReadOnly = _ro; }

Isn't this asking for trouble, i.e., an inline virtual method?

Paulo