- Mar 19, 2018
-
-
- Mar 05, 2018
-
-
Georges Basile Stavracas Neto authored
This grants full powers to consumers of this API. Long term, though, we'll remove this provider and make clients handle auth themselves. Thus, this is a short-term fix.
-
- Feb 26, 2018
-
-
Administrator authored
-
- Jan 27, 2018
-
-
- Dec 31, 2017
-
-
Matej Urbančič authored
-
Matej Urbančič authored
-
- Dec 14, 2017
-
-
Debarshi Ray authored
-
Debarshi Ray authored
-
- Dec 13, 2017
-
-
- Dec 08, 2017
-
-
- Dec 05, 2017
-
-
- Nov 25, 2017
-
-
Aurimas Černius authored
-
- Nov 17, 2017
-
-
- Nov 15, 2017
-
-
- Nov 11, 2017
-
-
- Nov 10, 2017
-
-
- Nov 08, 2017
-
-
Marek Cernocky authored
-
- Nov 07, 2017
-
-
-
-
Piotr Drąg authored
-
Debarshi Ray authored
So far, realmd was being merely used to auto-discover a list of realms, and validate whether the user entered a valid realm or not before prompting for credentials. This sounds nice in theory but, in practice, it is far too fragile. For example, realmd gets frequently tripped up by Fedora's Kerberos setup which uses _kerberos._tcp instead of the more common _kerberos._udp SRV record. Secondly, the realm validation isn't snappy and can take a while, so the user has to wait a bit before she is able to enter her credentials. The current UI is also odd because it presents separate fields for entering the realm and the username. Most people are used to entering username@realm or username@REALM because of their familiarity with email addresses and kinit. Therefore, the account addition UI has now been re-written without realmd. There is only one field for the user to enter her principal, and the realm is automatically capitalized if it isn't. It entirely relies on libkrb5.so to validate the principal during the initial sign-in once the credentials have been entered. https://bugzilla.gnome.org/show_bug.cgi?id=789187
-
-
Debarshi Ray authored
-
-
- Nov 06, 2017
-
-
Kjartan Maraas authored
-
- Oct 14, 2017
-
-
- Oct 05, 2017
-
-
- Oct 03, 2017
-
-
Debarshi Ray authored
-
Debarshi Ray authored
The binaries corresponding to the D-Bus services are not installed when the backend is disabled, so these service files are useless.
-
- Sep 30, 2017
-
-
Piotr Drąg authored
-
- Sep 22, 2017
-
-
- Sep 18, 2017
-
-
- Sep 14, 2017
-
-
Alexander Shopov authored
-
- Sep 13, 2017
-
-
Debarshi Ray authored
-
- Sep 12, 2017
-
-
- Sep 10, 2017
-
-
- Sep 09, 2017
-
-
Inaki Larranaga Murgoitio authored
-