GThread struct now hidden
- From: Murray Cumming <murrayc murrayc com>
- To: gtk-devel-list <gtk-devel-list gnome org>
- Cc: Ryan Lortie <desrt desrt ca>, gtkmm-list <gtkmm-list gnome org>
- Subject: GThread struct now hidden
- Date: Thu, 13 Oct 2011 12:48:26 +0200
This change in glib master does indeed break glibmm:
http://git.gnome.org/browse/glib/commit/?id=d904612100120d12126f1a6623a106d8a5b02fa6
Unless it's really really necessary, it would be great if you would not
do this, or if you could leave it until a glib ABI break.
Glib::Thread has a static (but private) GThread member instance:
http://git.gnome.org/browse/glibmm/tree/glib/src/thread.hg#n285
so obviously glibmm no longer builds.
We do this so you can cast a Glib::Thread* directly to a GThread*. We
probably did this (in 2002?) because we have no way to associate our C++
instance with the C instance, because GThread is not a GObject. I doubt
that we can find a way around that, though I have not tried.
We also access the struct fields (joinable, priority) directly in simple
getter methods, but I don't know if anybody really uses them. We might
get away with breaking their functionality if we can't find
replacements:
http://git.gnome.org/browse/glibmm/tree/glib/src/thread.ccg#n120
http://git.gnome.org/browse/glibmm/tree/glib/src/thread.ccg#n135
At the least, we should probably deprecate those methods.
I have no objection to deprecating our entire Glib::Thread API, if we
can replace it with a more-correct Glib::Thread2 API, but we'll be in
trouble if user applications (built with glibmm <= 2.30) start crashing
just because the user installed glibmm 2.32.
--
Murray Cumming
murrayc@murrayc com
www.murrayc.com
www.openismus.com
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]