GNOME Bugzilla – Bug 708336
Being able to use GDM via the keyboard (no mouse)
Last modified: 2015-04-16 08:33:13 UTC
I recently crashed my mouse and can not login into GNOME. It seems highly impractical not to use GDM options with the keyboard (like selecting user). My solution was to buy a new mouse, but I consider it very important to use GDM via the keyboard. Thanks
(In reply to comment #0) I've done some testing, and I've noticed that the keyboard works sometimes. When not working, the user list is not highlighted, and the keys as "TAB" or the arrow keys have no effect. However, if the user is selected with the mouse, the keyboard if it works when you type the password (also from time to select a user with the mouse, the keyboard begins to function normally). To consider problems in connecting peripherals to my desktop computer, I tried on a laptop, with the same results. fedora 19 32 bits GNOME 3.8.4 GDM (gdm-1:3.8.4-2.fc19) I could not identify the problem, but only occurs in some occasions (rebooting the system or by logging), in which any object is highlighted.
hmm, works fine here, with 3.9.92
(In reply to comment #2) > hmm, works fine here, with 3.9.92 I have identified the situation that causes the problem. The problem occurs after the screen is locked when not yet logged on. From what I see only occurs in the logon screen. I have not had chance to try on fedora rawhide, but the problem exists in GNOME 3.8.
I think I've seen this too. You also can't use Esc to clear the shield when GDM hits inactivity timeout, IIRC.
Has anything happened with this? Using gdm-3.10.2, I find that I can dismiss the shield with the enter key, but then no keys will select the only user available. This is a regression from 3.8 where enter would automatically select the first/only user in the list. This is entirely reproducible, and as pointed out demands a working mouse to be able to login to gnome 3.10.2. This might also relate to bug 709363 and bug 709885.
In 3.16 not reproduce the problem.
I cannot reproduce this either, so let's assume the issue got fixed at some point.