Bug 234140 - KNotify spams .xsession-errors; ignores kdebugdialog's "Disable all debug output"
Summary: KNotify spams .xsession-errors; ignores kdebugdialog's "Disable all debug out...
Status: RESOLVED FIXED
Alias: None
Product: kdelibs
Classification: Frameworks and Libraries
Component: general (show other bugs)
Version: 4.4
Platform: openSUSE Unspecified
: NOR normal
Target Milestone: ---
Assignee: kdelibs bugs
URL:
Keywords:
Depends on:
Blocks: 325889
  Show dependency treegraph
 
Reported: 2010-04-12 11:03 UTC by Tristan Miller
Modified: 2020-09-28 22:41 UTC (History)
9 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments
Sample .xsession-errors file as requested in Comment #2 (23.89 KB, application/x-bzip2)
2010-08-12 17:18 UTC, Tristan Miller
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Tristan Miller 2010-04-12 11:03:31 UTC
Version:            (using KDE 4.4.2)
Installed from:    openSUSE RPMs

KNotify outputs an awful lot of messages to ~/.xsession-errors which aren't really error messages, but rather what appears to be generic debugging info.  It shouldn't do this by default.  It outputs these messages even when "Disable all debug output" is checked in kdebugdialog.

See also Bug 227697.
Comment 1 Christoph Feck 2010-06-17 15:51:02 UTC
Could you provide an output similar to bug 227697 comment #7 so that it can be seen which messages from KNotify cannot be suppressed?
Comment 2 Tristan Miller 2010-08-12 17:18:36 UTC
Created attachment 50069 [details]
Sample .xsession-errors file as requested in Comment #2
Comment 3 Christoph Feck 2010-08-12 18:25:31 UTC
Thanks for the information. There are indeed messages from KNotify, but they account only for 892 out of 12789 lines (7%). Will be fixed.

The majority of messages seem to come from Akonadi, so I would suggest to report a similar bug to Akonadi developers, and link to this attachment.

Additionally, many messages come from bad .desktop files for YaST and Wine detected by kbuildsycoca4. Those errors could be filed downstream. Of course the messages should not appear when debug output is disabled.

As for the "CLIENT:" stuff, please report to Kopete developers.
Comment 4 Florian Heimgaertner 2010-09-16 22:40:41 UTC
Same problem here with KDE 4.4.5 (Debian squeeze). My ~/.xsession-errors is full of debug messages from knotify and kglobalaccel. Every use of KDE hotkeys is logged although debug output is disabled in kdebugdialog.
Comment 5 Christoph Feck 2010-10-10 01:11:19 UTC
Could you please check if you have a [knotify] section in .kde/share/config/kdebugrc ?
Comment 6 Tristan Miller 2010-10-10 11:19:20 UTC
Yes, I do have a [knotify] section.
Comment 7 nick 2010-11-17 01:43:59 UTC
I can confirm this with 4.4.5.
It even logs a line for every single keypress individually during an alt-f2. Gross.

The .xsession-errors-filling-up problem has gone on *way* to long--years. No viable work-a-round has surfaced either. Something still replaces a symlink to /dev/null with a regular file at login no matter what I try. I've grepped sources and cannot find who is doing it. I've commented out the part in /etc/X11/Session/XSession that creates it--no help there either.

I have to log out and log back in weekly to get rid of it--in under a week it reaches nearly 2GB, and everything becomes noticeably slower.
Comment 8 quazgar 2011-02-20 23:54:33 UTC
I can also confirm this behaviout with 4.4.5 on Gentoo, what's the state on 4.5 / 4.6?
Comment 9 nick 2011-02-21 00:01:12 UTC
The only work-a-round I've come up w/is to edit /etc/X11/startDM.sh and change:
(gentoo)

${NAME:+--name} ${NAME} ${PIDFILE:+--pidfile} ${PIDFILE} ${START_STOP_ARGS} || \

to:

${NAME:+--name} ${NAME} ${PIDFILE:+--pidfile} ${PIDFILE} ${START_STOP_ARGS} -- -error /dev/null || \

(add -- -error /dev/null)

This causes *all* of kdm + spawned processes output to go to dev null.
Not ideal, but it's a sledge-hammer approach that lets me work.

I have not been able to test 4.5 or 4.6--I'm waiting for kdepim to get their stuff together. As it is, kmail's imap support in 4.4.5 is very flakey so I really hope they get a new, polished kdepim in 4.6.1...
Comment 10 James J. Ramsey 2011-05-30 22:18:12 UTC
FYI, I had a problem with an overly large .xsession-errors, and in my case, the culprit seems to have been an errant panel applet. I had added the LCD weather applet, which stopped displaying itself on the panel but was still checkmarked as installed when I went to add it again to the panel. This led to messages like the following,

QPainter::pen: Painter not active
QPainter::setPen: Painter not active
QPainter::opacity: Painter not active
QPainter::brush: Painter not active
QPainter::pen: Painter not active
QPainter::setPen: Painter not active
QPainter::setOpacity: Painter not active
QPainter::drawPath: Painter not active
QPainter::setPen: Painter not active
QPainter::setOpacity: Painter not active
QPainter::setBrush: Painter not active
QPainter::setPen: Painter not active
QPainter::setWorldTransform: Painter not active
QPainter::setWorldTransform: Painter not active
QPainter::restore: Unbalanced save/restore

to appear over and over. When I saved a copy of .xsession-errors before logging out, my copy was about 55 MB in size and had 1,424,475 lines in it.
Comment 11 David Faure 2011-06-21 08:46:16 UTC
QPainter warnings are directly from Qt, so they do ignore the KDE setting "disable all debug output" indeed. Something which we should work on for Qt 5...

Meanwhile, if this is a LCD weather applet bug, please file a separate bug report for the LCD weather applet guys to see it and fix it.
Comment 12 David Faure 2011-12-21 19:16:15 UTC
Comment 3 says akonadi was responsible was most of the much debug output in release mode: confirmed, and fixed today.
Comment 13 dolgener 2014-08-21 12:55:17 UTC
In general, some possibility to "blacklist" certain messages (which are displayed frequently and/or without relevance to the user) 
a) from popping up or 
b) from being queued into the message list at all
would be great!
Comment 14 Nate Graham 2020-09-28 22:41:20 UTC
Yep, this seems to be fixed today.