Active discussions

Name ChangeEdit

There's a [by whom?] tag on the sentence "Since version 0.105, released in 2015[2] the name of the project was changed[by whom?] from PolicyKit to polkit [..]". This seems completely unnecessary to me. It is obvious that when a project changes its name, the name was changed "by" the project. There's no good way to ascribe a group decision like that to any individual, nor do I see what the point would be of doing so. If there's no arguments against it, I'm going to remove that tag.

Either the name change is not universal or there is a fork. Debian has policykit-1 and no polkit. John Talbut (talk) 14:43, 13 April 2018 (UTC)
 Debian dosn't ship updates that break backward compatibility. So there is a "grave"-Bug in PolKit that it breaks backward compatibility. As long as there is no solution for this problem like migration scripts debian will likely stay on the old PolicyKit version. IMHO it dosn't hurt. Changes where appreciated mostly by the devs themselves not so much by most users or admins who have to migrate all their scripts and have to learn another not very easy to (securely) master language. --2001:7C0:3006:301D:921B:EFF:FEE0:7EBB (talk) 09:23, 7 August 2020 (UTC)


This desparately needs expansion with more details, as well as a tie-in to other related articles and how-to links. It is missing a connection to basic authorization and permissions articles. - KitchM (talk) 02:54, 7 March 2010 (UTC)

Linus Torvalds about polkitEdit

"Whoever moron thought that it's "good security" to require the root password for everyday things like this is mentally diseased." — Preceding unsigned comment added by (talk) 19:34, 9 November 2012 (UTC)

It's definitely a design intended for easy desktop use by beginner users, and possibly partly inspired from systems like Apple/NetBSD kauth and other MAC (Mandatory Access Control) and java policies, but limited to a higher level scope. Seasoned unix sysadmins and systems programmers may consider this pervasive (like sudo, extensive setuid usage, and to some extent, dbus, RPC with privilege separation, xrandr, etc), to be a type of "controled privilege escalation" or "controled security hole", even if it's poked though some preferences (which might also be security-suboptimal using the distribution-supplied defaults). This is information which could be added, but we also need to find references and properly present this in an objective enough way that does not sound too much like a rant :) (talk) 02:04, 29 March 2013 (UTC)
Return to "Polkit" page.