On Thu, Oct 01, 2009 at 14:57:19 +0200, Jiří Zárevúcky wrote:
> On Wed, Sep 30, 2009 at 7:46 PM, Jan Hudec <[email protected]> wrote:
> > On Wed, Sep 30, 2009 at 18:58:54 +0200, Philipp Zabel wrote:
> >> I'd like to subclass GLib.BufferedInputStream.
> >> Vala doesn't like the following construct:
> >>
> >> class MyStream : GLib.BufferedInputStream {
> >> MyStream (GLib.InputStream base_stream) {
> >> base (base_stream);
> >> }
> >> }
> >>
> >> $ valac mystream.vala --pkg gio-2.0
> >> mystream.vala:3.3-3.20: error: chain up to
> >> `GLib.BufferedInputStream..new' not supported
> >> base (base_stream);
> >> ^^^^^^^^^^^^^^^^^^
> >> Compilation failed: 1 error(s), 0 warning(s)
> >>
> >> Am I doing it wrong? BufferedInputStream has a constructor that takes
> >> one GLib.InputStream.
> >
> > Short story:
> > ------------
> >
> > MyStream (GLib.InputStream _base_stream) {
> > base_stream = _base_stream;
> > }
> >
> > Long story:
> > -----------
> >
> > [...]
> > --
> > Jan 'Bulb' Hudec
> > <[email protected]>
>
> I take it this has changed recently?
It was like this from the begining. It must have been like this, because it's
how GObject library is designed.
> Last time I checked constructor chaining worked flawlessly. Is there
> any documentation on this new behavior?
It does work. Am I saying it does not? It's just that you can't use
non-default constructors of non-vala classes, because they are not really
there. And you never could, because they were never there.
--
Jan 'Bulb' Hudec <[email protected]>
_______________________________________________
Vala-list mailing list
[email protected]
http://mail.gnome.org/mailman/listinfo/vala-list
