Konq/E CVS Head: past the vpath problem

[Keith Kyzivat]

Ok, I've found what looks like a real bug now.

    html_objectimpl.cpp: In method `class KJavaApplet * 
DOM::HTMLAppletElementImpl::applet() const':
    html_objectimpl.cpp:153: parse error before `*'

Looking in html_objectimpl.cpp:45-48, I see this:
    #ifndef Q_WS_QWS // We don't have Java in Qt Embedded
    #include "java/kjavaappletwidget.h"
    #include "java/kjavaappletcontext.h"
    #endif

html_objectimpl.cpp:148-154:
    KJavaApplet* HTMLAppletElementImpl::applet() const
    {
        if (!m_render || !m_render->isApplet())
            return 0L;

        return 
static_cast<KJavaAppletWidget*>(static_cast<RenderApplet*>(m_render)->widget())->applet();
    }


Luciano, have you tried this with Qt/Embedded? or were your tests with 
Qt only?
It's looking like you tried it with Qt only.


So, since Qt/E doesn't support Java, I simply #ifndef'ed the return and 
in the case of Q_WS_QWS, put in a null return.

Bear with me a little -- I'm still getting into the swing of doing 
actual code changes to OS projects, what is politically correct,
when to talk so you don't bother the core developers time, etc.

Ok, next error.. unsure how to proceed from here.  This one makes me 
quite sure that you've been using Qt and not Qt/e...
Qt/E doesn't support rtti (defines -fno-rtti), and I've got a place 
where a dynamic_cast is used (rightfully so).

    khtmlview.cpp: In method `void KHTMLView::closeChildDialogs()':
    khtmlview.cpp:473: taking dynamic typeid of object with -fno-rtti

Not sure how to proceed from here.  Is there some sort of Hashed class 
ID in a base class that can be used for identification?


~Keith Kyzivat