Bug 359661

Summary: Digital clock is not updated consistently
Product: [Plasma] plasmashell Reporter: Fabio Coatti <fabio.coatti>
Component: Digital ClockAssignee: Martin Klapetek <mklapetek>
Status: RESOLVED DUPLICATE    
Severity: normal CC: bugs.kde.org, donjajo4all, kde, kde, kde, plasma-bugs
Priority: NOR    
Version: 5.7.1   
Target Milestone: 1.0   
Platform: Gentoo Packages   
OS: Linux   
Latest Commit: Version Fixed In:
Sentry Crash Report:

Description Fabio Coatti 2016-02-22 06:51:20 UTC
The time shown by digital clock in default panel (lower right bottom) is updated with some delay or even jumping some seconds. Say, from :49 it goes to :51. I've seen that also for minutes, when I configured it to do not show seconds.
It seems to have this behaviour only in 5.5.4, but I can't be sure of this. 
My usual config is to show seconds, timezone and long date format.
Comment 1 Kai Uwe Broulik 2016-02-24 14:20:57 UTC
When you open Kickoff ("the start menu") and hover entries there, do you also only have every second item highlighted as you move around? If so, this is a driver issue.
Comment 2 Fabio Coatti 2016-02-24 14:24:23 UTC
Yep, exactly so!

the higlight happens only on second item, good catch. What driver are you referring to? video card? something in mesa layer?

Many thanks for the answer!
Comment 3 Jajo 2016-02-26 10:10:55 UTC
Same issue here
I use Debian Stretch and Plasma 5
Comment 4 Fabio Coatti 2016-07-14 07:49:41 UTC
After a while, the issue is reappeared. I don't think is related to xf86-intel driver (I'm on 2.99.917 and I've also tried to downgrade it).

Now I'm on plasma 5.7.1 (same behaviour on 5.7.0), gentoo packages. I can provide all information required or try different versions of packages, just let me know.
Comment 5 Erik Quaeghebeur 2016-07-25 21:34:47 UTC
I can confirm this behavior on plasma 5.5.5. It appears it's also been reported elsewhere: https://bugzilla.redhat.com/show_bug.cgi?id=1307178
Comment 6 kde 2016-07-26 23:15:39 UTC
I can confirm this behavior on Plasma 5.6.5 (using backports PPA on Kubuntu 16.04). The clock widget only seems to update once every two seconds. Occasionally it'll show two consecutive seconds. My usual display is 24-hour time with seconds and long date, e.g.

16:10:26
Tuesday, July 26, 2016

I thought maybe it might have something to do with the fact that I normally have two clock widgets, one at the right edge of the bottom panel on each of my two monitors, but I removed one of the clocks and the other clock still only updates every ~2 seconds or so.

Also, Plasmashell is currently using about 30% CPU according to GNU top. (By contrast, on my old 14.04 machine, plasma-desktop is currently using about 0.7% CPU.) I killed plasmashell (with "kdequitapp5 plasmashell") and then ran plasmashell & again, and now it's using around 1% CPU, and the clock widget is updating every second.

So it looks like it's not a bug with the clock per se, but the fact that Plasmashell eventually decides to just start chewing up tons of CPU for no apparent reason.
Comment 7 Erik Quaeghebeur 2016-07-27 09:04:11 UTC
(I'm now at Plasma 5.6.5.1)

(In reply to kde from comment #6)
>
> Also, Plasmashell is currently using about 30% CPU according to GNU top. [...]
> I killed plasmashell (with "kdequitapp5 plasmashell") and then
> ran plasmashell & again, and now it's using around 1% CPU, and the clock
> widget is updating every second.
> 
> So it looks like it's not a bug with the clock per se, but the fact that
> Plasmashell eventually decides to just start chewing up tons of CPU for no
> apparent reason.

I get the second-skipping behavior even when plasma eats no discernible amount of cpu resources. So although this may also cause this issue, or exacerbate it, it cannot be the root cause.
Comment 8 Erik Quaeghebeur 2016-07-27 09:09:46 UTC
(In reply to Fabio Coatti from comment #4)
>
> xf86-intel driver (I'm on 2.99.917 [...]).

N.B.: I'm using the same driver and also have the entry-skipping behavior.

Do all affected people use the (this?) intel driver?
Comment 9 kde 2016-07-27 14:22:54 UTC
(In reply to Erik Quaeghebeur from comment #8)
> (In reply to Fabio Coatti from comment #4)
> > xf86-intel driver (I'm on 2.99.917 [...]).
> 
> N.B.: I'm using the same driver and also have the entry-skipping behavior.
> Do all affected people use the (this?) intel driver?

Yeah, I'm using that same driver too.
Comment 10 David Edmundson 2016-08-25 01:47:18 UTC

*** This bug has been marked as a duplicate of bug 353983 ***
Comment 11 Fabio Coatti 2016-08-25 06:54:15 UTC
I tried to remove xf86-intel driver completely, relying on kms setup, and the issue disappeared. I don't really know what are the drawbacks, maybe the grpahics (not really sure, though) are a bit slower but for my usage it is fine.
Comment 12 Erik Quaeghebeur 2016-08-25 19:33:41 UTC
(In reply to David Edmundson from comment #10) 
> 
> *** This bug has been marked as a duplicate of bug 353983 ***

David, I don't see how this is a duplicate. Can you explain?