On Saturday 16 December 2006 00:46, Phil Endecott wrote:
> >> Maybe there is some subtle bug in there.
>
> OK, how about this. In ++DFB/include/++dfb.h, there is what looks like
> a copy constructor:
>
> IPPAny(const IMPLEMENTINGCLASS &other) {
>
> I don't think that's ever being called - debug statements added there
> don't print.
hi phil,
thanks for pointing out this bug. i committed a fix to cvs just a min ago.
the next days, i'll have a closer look on the sources again. i think i still
wanted to get a lot of stuff to be const. after that i'll also increase the
version level according to DirectFB itself.
> I think that's because it has the wrong parameter type to be recognised
> as a copy constructor. The copy constructor for class X must be
> X(const X &), i.e. in this case:
>
> IPPAny(const IPPAny<IMPLEMENTINGCLASS,IPPAny_C> &other) {
>
within a template declaration you just use the class name.
IPPAny(const IPPAny &other);
therefore that is the correct version.
as we use all IDirectFB.... classes just as an encapsulation for surfaces we
dont have to
have a copy ctor for each class seperately. the one within the template works
now and
is called. i used your example program to test that, it works now.
(within an x11 window, i am not using fullscreen dfb on the pc)
> Making that change, my test programs at least get a bit further than
> they used to.
> Any comments? (What about the operator= in the same file?)
>
>
> Phil.
>
best regards
marcel
_______________________________________________
directfb-users mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
