GNOME Bugzilla – Bug 701620
unknown signal name: changed
Last modified: 2014-03-01 01:46:19 UTC
I'm getting a rather large amount of terminal output every time a file is opened from a diffdir (and other times...) Traceback (most recent call last):
+ Trace 232016
entry.connect("changed", lambda *args: self.emit("changed"))
I suspect it may be OS specific - I'm running ArchLinux when receiving this one. I can't discern what exactly this is trying to do unfortunately so couldn't debug further =(
Sadly, I can't reproduce this. I suspect it might be GLib + GTK-version-dependent... can you tell me what you're running? This *did* change in 1.7.3 because HistoryEntry was split and gained a parent class. It looks to me like maybe the parent signals aren't being added to the child class for some reason. Alternatively, this could have something to do with the adding-interfaces-after-class-creation API breakage in... glib 2.36?
Created attachment 246477 [details] [review] Fix console messages re HistoryFileEntry
Gnome 3.8.3 glib2 2.36.2-1 lib32-glib2 2.36.2-1 clutter-gtk 1.4.4-1 gtk-engines 2.21.0-1 gtk2 2.24.18-1 gtk3 3.8.2-1 Arch is generally pretty bleeding edge. I suspect your latter suggestion is most likely? My GTK fu is weak - where should this interface be defined? Unfortunately grep'ing for changed isn't exactly useful - and no methods seem to be named 'changed'. I did notice that __gsignals__ at the top of the class has 'activate', but no 'changed' - should it be there? When added that stops the messages appearing *shrug* Patch attached, but I don't know enough to know if it's a reasonable fix or not :\
I'm not really comfortable with that fix... I mean I think you're probably right, but I can't tell whether or not it's the right fix. I'll see if I can get an environment to reproduce the problem and test the fix (and other possibilities). If the problem is what I think it is, then it's the fact that we're inheriting from the gtk.Editable interface.
Did more digging. I'm certain that this is the problem. Apparently it's occurred upstream, but forwarding bugs is hard or something: https://bugs.launchpad.net/ubuntu/+source/meld/+bug/1103170 There's an attempt to fix this in pygobject 2.28, but there's a failure case reported: https://bugzilla.gnome.org/show_bug.cgi?id=694108 I think we dodge that case, but only through missing super() calls and not through any judicious whatever. I'm open to any way we can fix this in Meld, but I'm not sure whether it's possible... so that's nice.
Since this is against Meld 1.8.x and we can't really fix it anyway, I'm not sure there's much point keeping this bug open. It's annoying, but... yeah. Thanks anyway.