[Wayland-bugs] [Bug 700166] glade fails to build introspection with Gtk 3.9

glade (bugzilla.gnome.org) bugzilla at gnome.org
Mon May 13 12:18:00 PDT 2013


https://bugzilla.gnome.org/show_bug.cgi?id=700166
  glade | general | unspecified

--- Comment #11 from Tristan Van Berkom <tristanvb at openismus.com> 2013-05-13 19:17:55 UTC ---
(In reply to comment #9)
> (In reply to comment #7)
> 
> > Sure it can.
> 
> You're correct; sorry, I went off the rails briefly.
> 
> > Only the GIR will not contain an enum with the stock ids listed as
> > enum values (which as I mentioned, is not important anyway).
> > 
> > It's a plausible workaround afaics.
> 
> Yeah, I guess.  But it's still quite ugly to have the situation where the GType
> for GladeStock can only be accessed after gtk_init() has been invoked.

And I definitely disagree with this, while yes... Glade could be refactored
to work differently wrt editing stock strings... this exact statement is
IMO incorrect.

Glade depends on GTK+, why does it make any sense that any Glade code
should be callable before gtk_init() ? That in itself is not ugly (there
may be ugly things in Glade, but that in itself is not one of them).

This is actually, the whole root of this debate IMO, of course I could be
just implementing the workaround (which I'll probably end up doing anyway),
the whole interesting part of this discussion is that introspection somehow
thinks it's OK to access code in libraries without initializing those
libraries.

Sure, maybe this is something that we should end up living with ?

Maybe we should state that type initializers and class initializers
somehow don't count as API ? and document/enforce that as a rule ?

Maybe we should document/enforce that libraries should have ways to
initialize themselves in environments like build servers, where some
D-Bus services are not available, where there might not be a DISPLAY ?

Whichever it is, we should look at it.

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.


More information about the Wayland-bugs mailing list