Re: let's split the systemd binary package
On 25/10/13 14:39, Paul Tagliamonte wrote:
> On Fri, Oct 25, 2013 at 01:40:55PM +0200, Olav Vitters wrote:
>> ... a choice between something greatly supported (logind) vs
>> something abandoned (ConsoleKit).
> Since the project (on the whole) is fairly divided, I don't think
> we should trivialize this to "actively developed" vs "cruft" at
> this stage.
The fact remains that systemd *is* actively developed, and ConsoleKit
isn't. If GNOME upstream refuse to use an unmaintained framework for
session-tracking (which seems entirely reasonable to me, tbh), then
the options seem to go something like this:
* someone who doesn't want logind does the work to maintain a
session-tracking framework with the features GNOME needs; or
* someone who doesn't want logind turns the features that need
session-tracking into a no-op on non-logind systems, and deals
with the necessary Depends -> Recommends downgrades and the
"help, features X and Y don't work!" bug reports that result;
* GNOME continues to depend on logind.
"Session tracking" includes suspending/hibernating, because logind has
a mechanism to let apps delay suspend, which is necessary for things
like closing the inherent race condition in "lock the screensaver when
we suspend... oh, oops, it didn't get scheduled until after we
resumed, so the old screen contents are still visible for a moment
when you open your laptop".
My understanding is that Debian's GNOME maintainers aren't keen on the
second of those options, because as a small team maintaining a large
number of non-technical-user-facing packages, they already have far
more bug reports than they can deal with.