[Webkit-unassigned] [Bug 29678] Conditionally guard cursor code with !QT_NO_CURSOR

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=29678





--- Comment #3 from Norbert Leser <norbert.leser at nokia.com>  2009-09-23 06:48:45 PDT ---
(In reply to comment #2)
> (From update of attachment 39991 [details])
> 
> > Index: WebKit/qt/Api/qwebgraphicsitem.cpp
> > ===================================================================
> > --- WebKit/qt/Api/qwebgraphicsitem.cpp	(revision 48670)
> > +++ WebKit/qt/Api/qwebgraphicsitem.cpp	(working copy)
> > @@ -95,12 +95,16 @@ void QWebGraphicsItemPrivate::update(con
> >  
> >  QCursor QWebGraphicsItemPrivate::cursor() const
> >  {
> > +#ifndef QT_NO_CURSOR
> >      return q->cursor();
> > +#endif
> >  }
> >  
> >  void QWebGraphicsItemPrivate::updateCursor(const QCursor& cursor)
> >  {
> > +#ifndef QT_NO_CURSOR
> >      q->setCursor(cursor);
> > +#endif
> >  }
> 
> The patch looks good to me, expect for these two hunks. Instead of placing the
> #ifdef into the function it should be placed
> around the definition/declaration. In other words: If QT_NO_CURSOR is defined,
> then those functions not not be declared
> in the first place, similar to the guards around QWidget::setCursor()/cursor()
> in qwidget.h.

That is your call, or whoever implemented the classes/functions. In the case of
conditionally guarding the functions itself, you may end up with a multitude of
conditional statements in the client code. Otherwise, it would be transparent
(though, perhaps unexpected behavior - i.e., updateCursor etc would do nothing,
unless there is another default implementation).

Anyway, could you or the implementer of QWebGraphicsItemPrivate,
QWebViewPrivate, and QWebPageClient take it from here? - Thanks.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.