Bug 12038 - xscreensaver hack continues in background
Summary: xscreensaver hack continues in background
Status: RESOLVED INVALID
Alias: None
Product: Mageia
Classification: Unclassified
Component: RPM Packages (show other bugs)
Version: 3
Hardware: i586 Linux
Priority: Normal normal
Target Milestone: ---
Assignee: Mageia Bug Squad
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2013-12-18 17:20 CET by Doug Laidlaw
Modified: 2014-01-28 15:26 CET (History)
1 user (show)

See Also:
Source RPM: xscreensaver-5.21-1.mga3
CVE:
Status comment:


Attachments

Description Doug Laidlaw 2013-12-18 17:20:39 CET
Description of problem:
Running Mga3 with XFCE and all updates.
Since the updates about Dec 17, when I stop xscreensaver and bring back my desktop, the executable for the hack continues in background, using 10-15% CPU. It won't respond to a kill command.  I have to use "kill -9"

At the moment, ps shows:

doug     11273 10.4  1.6  80908 68496 ?        RN   03:06   1:07 glslideshow -root


Version-Release number of selected component (if applicable):
5.21-1.

How reproducible:
Every time the screensaver is stopped to bring the desktop back, whether after normal scredensaving or using startsaver to start it manually.  Lock screen is off.

Steps to Reproduce:
      As Above.
1.
2.
3.


Reproducible: 

Steps to Reproduce:
Comment 1 Doug Laidlaw 2013-12-18 22:00:31 CET
After shutting down overnight, this bug appears to have vanished.  I will watch it for a few days.  Last night, at one point, there were two hacks running in the background.
Comment 2 Doug Laidlaw 2013-12-18 22:04:51 CET
a different topic.  The yes command is giving 

yes: standard output: Broken pipe
yes: write error

but the command following is still executed.  Probably should be a separate bug, but somebody in programming will no doubt come up against it.
Comment 3 David Walser 2013-12-19 21:56:52 CET
In what context does the yes command give an error?  Just running it in a terminal works fine.

As for xscreensaver, if the problem does come back, it'd be interesting to know if 5.26 fixes it.  You could try rebuilding it from the source RPM with 5.26 and installing an updated RPM.  If it does fix an issue for you, we could backport it as an update.

I'm going to close this as INVALID for now, but feel free to post any additional comments or reopen it if there's a reproducible issue.

Status: NEW => RESOLVED
CC: (none) => luigiwalser
Resolution: (none) => FIXED

Comment 4 David Walser 2013-12-19 21:57:11 CET
Whoops, INVALID not FIXED.

Resolution: FIXED => INVALID

Comment 5 Doug Laidlaw 2013-12-19 23:59:29 CET
INVALID is fine by me.  It hasn't come back.

Both these problems arose after the latest round of updates.  There is another post on alt.os.linux.mageia about a problem with kaffeine disappearing after a cold start.

Just running yes in a terminal works fine for me too.

I run bashpodder.  I added a line to copy my playlist to a backup as follows:

yes | cp 2012-10-20/podnew.m3u ./podnew.bak

That now gives me the quoted error message.  The command that follows, still runs to completion.  Instances I found on the Web are identical in this regard, but are in foreign contexts: BSD, a specific program.  Last night I saw it in the output of something else.  It sounds like it's just me and a nuisance thing anyway.
Comment 6 David Walser 2013-12-20 00:35:59 CET
Ahh, you get that output because cp stops reading from standard input after reading the first 'y' and then the yes process gets killed.  Doesn't look like there's a way to tell yes how many iterations to run.  I'd think it'd make more sense to just use cp -f, then you wouldn't need yes.
Comment 7 Doug Laidlaw 2013-12-20 03:11:02 CET
Thanks for the tip.  That cured it.
Comment 8 Doug Laidlaw 2014-01-21 14:32:08 CET
The xscreensaver problem is back.  I had to reinstall Cauldron again because I couldn't get beyond a "dropping to maintenance" prompt.

Also back is Bug 11666, which vanished in the same way.  I probably should be running the official release (or another distro.)

Status: RESOLVED => REOPENED
Resolution: INVALID => (none)

Comment 9 Doug Laidlaw 2014-01-21 14:34:13 CET
Sorry, this is using the 5.26 package David recommended.  Should I rebuild it?
Comment 10 David Walser 2014-01-21 14:58:42 CET
If you're using Cauldron, you shouldn't need to rebuild anything.  What exactly is the issue now?
Comment 11 Doug Laidlaw 2014-01-21 15:15:05 CET
It is exactly as in the Description.  There is no problem until a hack starts, either because the screensaver is activated, or because I select "Preview" in its config dialog.

The hack should exit when the desktop comes back up, but it keeps running in the background.  I run gkrellm, which shows Boinc as No.1 and the hack as No.2.  It doesn't seem to depend on which hack is involved, but I have not been able to check whether it is only on OpenGL ones.

I have looked at Bug 12230, which seems vaguely similar, but wasn't helpful.
Comment 12 David Walser 2014-01-21 17:19:27 CET
I ran xscreensaver-demo, it asked to start xscreensaver since it wasn't running (I was in IceWM), I chose GLSlideShow, hit Preview, saw the screensaver, clicked to make it go away, then closed xscreensaver-demo.  glslideshow did not persist in the process list.  I cannot reproduce.
Comment 13 Doug Laidlaw 2014-01-21 17:33:29 CET
So it looks as though it is only me.  It hasn't happened before the current Cauldron.  I didn't mention that this time, I deleted ~/.xscreensaver and re-created it, but I expect that it is happening elsewhere.

I am not sure that it ever really went away.  It no longer seems to take up as much CPU as before.  

There is something strange about this and Bug 11666 appearing together and going away together.  Perhaps a hardware problem?

I have another user (boinc.) I will see what happens there.  If that fixes it, I have a corrupted file somewhere.
Comment 14 Doug Laidlaw 2014-01-21 18:09:35 CET
Running it under boinc, the problem still occurs, but xscreensaver is as buggy as anything.  It may have been that I was in KDE, but it sounds as though I need to delete everything and do a first-time install, including my $HOME directory.
Comment 15 Doug Laidlaw 2014-01-27 15:30:05 CET
Cauldron just became unusable for the 500th time.  Aftert a reinstall, this problem seems to have gone away, but the GLX bug is back, so I can run only the "Themed" hack.
Comment 16 David Walser 2014-01-27 17:16:20 CET
OK, I'll close this bug.  If you're having video driver issues, please file a new bug.

Status: REOPENED => RESOLVED
Resolution: (none) => INVALID

Comment 17 Doug Laidlaw 2014-01-28 01:07:04 CET
Fair enough David.  I think that the gremlins have taken over.  I want to stay away from Bugzilla as much as possible.

After updating, I couldn't get a desktop.  I installed Mga3 on another partition, changed the Cauldron runlevel to 3 then rebooted into Cauldron with rebootin.

I finished up in a graphical desktop as normal, i.e. runlevel 5, back before Bug 12416 was fixed. A fix for the GLX came through about then.  I must have got the runlevel targets confused.  I have had enough of tearing my hair out over Cauldron.
Comment 18 David Walser 2014-01-28 01:10:02 CET
I hope Mageia 4 works out better for you.  Other than bugzilla, you may find other forms of support helpful too, such as the forums, the discuss mailing list, or IRC if you're so inclined.  Good luck.
Comment 19 Doug Laidlaw 2014-01-28 15:26:16 CET
I have joined Bug 12108.

Note You need to log in before you can comment on or make changes to this bug.