GNOME Bugzilla – Bug 674414
Cannot link Galaxy S II to computer with gnome-bluetooth 3.4.0
Last modified: 2012-07-02 10:25:32 UTC
See full bug report at https://bugzilla.redhat.com/show_bug.cgi?id=810615 Basically, it looks like gnome-bluetooth/applet/bluetooth-applet.c:confirm_request() is emitting the signal with a null name , which is causing the Bluetooth code in gnome-shell to produce a JS error- we could default name to <unknown> or something in gnome-shell, but probably it's better to handle things in gnome-bluetooth? $ rpm -q bluez gnome-shell gnome-bluetooth bluez-4.99-1.fc17.x86_64 gnome-shell-3.4.0-1.fc17.x86_64 gnome-bluetooth-3.4.0-1.fc17.x86_64
The right place (and Bastien will confirm, I think) to work around this is gnome-bluetooth, as it already has some hacks for specific devices. But I also see a critical trying to decode a GVariant, maybe it is a dbus failure down the stack?
Search for gnome-bluetooth bugs on redhat bugzilla in fedora 17 and you will find lots of folks reporting they can no longer pair anything (with no comments from any maintainers as if the whole issue is being ignored in the hopes that it will go away :-).
(In reply to comment #2) > Search for gnome-bluetooth bugs on redhat bugzilla in fedora 17 and you will > find lots of folks reporting they can no longer pair anything Which has nothing to do with this bug report. > (with no comments > from any maintainers as if the whole issue is being ignored in the hopes that > it will go away :-). Yeah, they might be too busy filing package updates.
I am also having this issue. What information can I provide to assist in debugging?
(In reply to comment #4) > I am also having this issue. What information can I provide to assist in > debugging? I doubt this is the problem you're seeing. Make sure you're using gnome-bluetooth 3.4.1 and file a separate bug.
Symptoms appear to be identical, down to the contents of .xsession-errors, using gnome-bluetooth 3.4.1. Should I still file a separate bug?
Created attachment 217802 [details] [review] Fix the parsing of the dbus return, so valid information is passed to the user
The pairing issues should be fixed with the above patch, given gnome-bluetooth falls back the Address in case there is no name it shouldn't really ever return a NULL. In case it does, either bluez broke or something else went horribly wrong then most likely the pairing attempt should be aborted either by gnome-bluetooth or gnome-shell.
FWIW, I don't think this actually fixes the pairing itself, but it fixes the error gnome-shell threw. The pairing problem itself was also fixed by Sjoerd in bug 679190, and was related to Simple Pairing (the "can you confirm the code" difference in the Fedora bug at https://bugzilla.redhat.com/show_bug.cgi?id=810615#c6 should have been the hint).