Re: segfault .....

["muppet" <scott asofyet org>]

Ross McFarland said:
> Bruno Boettcher said:
> > Ross McFarland said:
> > > that doesn't really look like it's a Gtk2 problem. Gtk2::Window->new is a
> > > trivial wrapper around gtk_window_new so it's unlikely the problems are
> > > stemming from that. it's dying pretty low down in gtk/gdk to boot.
> > >
> > > my only, unlikely, thought would be that maybe Gtk2->init (or use Gtk2
> > > '-init') was called before the call to Gtk2::Window->new?

on seeing the backtrace in the earlier message, that was my first thought ---
no gtk_init().


> > > the app is dying in
> > > settings related stuff which you probably aren't dealing with in your app.
> > hmm nope....
> > i am running this with the POE - wrapper....
> > all the graphical elements are build before the call to init is made....

that just doesn't work.  and, honestly, i don't understand why the gui has to
be built before gtk_init() because of POE --- how are they linked in any way?


> according to:
>     http://developer.gnome.org/doc/API/2.0/gtk/gtk-General.html#gtk-init
> "Call this function before using any other GTK+ functions in your GUI
> applications."
>
> i will be surprised if this fixes anything, but it could. on really old gtk
> there were some issues that required gtk_init to be called before just about
> anything could be done.

our test suite is full of examples of this, mostly related to the use of X
core fonts and lots of other little things.  since Glib and Gtk2 call
g_type_init() during bootstrapping, you won't see the type-system related
failures that come up in C programs which forget to call gtk_init() first.



-- 
muppet <scott at asofyet dot org>