GNOME Bugzilla – Bug 592127
Port to PolicyKit1
Last modified: 2009-09-01 09:31:39 UTC
We need to move to PolicyKit1 soon or later. If possible, I'd rather do that for 2.28 so that we stay consistent with the rest of the system. The new API is simpler and allow us to get rid of gst-polkit-button.{ch}. Attached is a first version of a patch.
Created attachment 140991 [details] [review] Migrate-to-PolicyKit1, v1
I haven't tried it out, or even read through in detail to see what is going on... but judging by the number of lines of code deleted, it looks like a good direction to go in :) I'd never heard of PolicyKit1 before today... what mailing list/rss feed was I supposed to subscribe to to find out about this?
(oh, I will have a better look at this later)
You can read the PolicyKit mailing list, in particular this message: http://lists.freedesktop.org/archives/polkit-devel/2009-July/000176.html Else you have the doc reference, but it's not always explicit on how to port form the previous version: http://hal.freedesktop.org/docs/polkit/index.html http://hal.freedesktop.org/docs/polkit-gtk/ We use polkit-gobject in the stb, and polkit-gtk in the gst. The funny thing is that we drop the PolicyKit dependency in the GUI, and only use one widget: PolkitLockButton. I'll try to commit my patches soon, they seem to be working. The problem is to get a release of the stb before that. Then I'd really like people to test it, in particular when it enters distributions. We're a little short in the timeframe, but there will be one pre-release left to revert the changes if we discover important issues. BTW, testing this from a distribution that does not include PolicyKit1 is a nightmare, I've spent hours on it. Better get a development release shipping it...
OK, I've pushed an updated patch as 517ee4c35559d1a368810fde2aefadf9ea570069, with and several other fixes. Please test as much as possible and report. Note that this requires the system-tools-backends 2.8 to work (configure will check that now).
So can this be closed as FIXED?
Sure, are you so worried about that? ;-) So far it does not seem to raise any bug, since I have seen not reports, especially from Ubuntu users, which seem quite reactive when a real problem appears (about 10 duplicate reports of a recently introduced crash). There's only one report that I'm trying to work out, but this kind of problems already occurred before the update, so... (for reference, it's https://bugs.launchpad.net/bugs/411533)
Yes I am. Because keeping stuff open that is supposed to be FIXED is not how a bugtracker works, plus we want to get rid of PolicyKit 0.9 completely for GNOME 2.28. :-)