Re: Crash processing g_object_new arguments

[Braden McDaniel <braden endoframe com>]

On Mon, 2009-11-09 at 09:32 +0100, Nicola Fontana wrote: 
> Il giorno Mon, 09 Nov 2009 03:24:04 -0500
> Braden McDaniel <braden endoframe com> ha scritto:
>
> > Clearly the "<Address ... out of bounds>" looks fishy; but I'm just
> > not seeing what could be a problem there; and I'm wondering if gdb
> > isn't just messing with me.  The call site for
> > openvrml_xembed_browser_new looks like this:
> >
> >         OpenvrmlXembedBrowser * const browser =
> >             OPENVRML_XEMBED_BROWSER(
> >                 g_object_new(OPENVRML_XEMBED_TYPE_BROWSER,
> >                              "control-host-proxy", host_proxy,
> >                              "control-host-name", host_name,
> >                              "dbus-thread-context",
> > dbus_thread_context, "expect-initial-stream", expect_initial_stream,
> >                              0));
>
> Change 0 to NULL: on 64 bit platforms 0 != NULL, hence the property list
> is not terminated.

Argh.  Of course.

0L would presumably have worked fine--until the code was compiled on a
platform where sizeof (long) != sizeof (void*).  (Hence Murray's
comment.)

Thanks, folks.

-- 
Braden McDaniel <braden endoframe com>