What | Removed | Added |
---|---|---|
CC | alarrosa@suse.com, mlin@suse.com | |
Assignee | fabian@ritter-vogt.de | bnc-team-screening@forge.provo.novell.com |
Summary | KDE:Frameworks5/plasma5-workspace: Bug Log-spamming "qt.qpa.xcb: QXcbConnection: XCB error"; patch available | Qt doesn't allow to disable logging of XCB errors |
Flags | needinfo?(alarrosa@suse.com), needinfo?(mlin@suse.com) |
(In reply to pgnd _ from comment #0) > Running > > lsb_release -rd > Description: openSUSE Leap 15.0 > Release: 15.0 Please report this bug in the right product then. > journactl fills with these > > Nov 09 21:21:28 pg.loc plasmashell[2876]: qt.qpa.xcb: QXcbConnection: XCB > error: 2 (BadValue), sequence: 6410, resource id: 62914566, major code: 142 > (Unknown), minor code: 3 > > lots > > journalctl | grep "qt.qpa.xcb: QXcbConnection: XCB error: 2 (BadValue), > sequence:" | wc -l > 47122 > > These/similar spamming log message have been reported for years, with > everybody pointing at someone else > > https://bugs.kde.org/show_bug.cgi?id=366421 > https://forum.qt.io/topic/55144/qxcbconnection-xcb-error-qwarning > https://bugreports.qt.io/browse/QTBUG-55167 > https://bugzilla.redhat.com/show_bug.cgi?id=1497564 > https://bugs.archlinux.org/task/57497 > > Finally, the last link suggests patches were made available for Fedora > > https://bugzilla.redhat.com/show_bug.cgi?id=1497564#c11 > > , i.e., at the distro. That's just a downstream bug which links to an upstream bugreport which I've already been tracking. It doesn't look like upstream is interested, for some reason. The real source is https://bugreports.qt.io/browse/QTBUG-55167 (which you linked already). > Here, on Opensuse, with > > rpm -q --whatprovides /usr/bin/plasmashell > plasma5-workspace-5.14.3-lp150.429.1.x86_64 > > (from, > http://download.opensuse.org/repositories/KDE:/Frameworks5/openSUSE_Leap_15. > 0) > > this > cat /etc/xdg/QtProject/qtlogging.ini > [Rules] > qt.qpa.xcb.xcberror=false > > has no effect on these messages. That has nothing to do with plasma, that's a Qt configuration option. The patch isn't a real fix either, it just makes it possible to hide the messages. IMO the messages are actually useful and shouldn't be hidden by default as they're actual errors. The patch is harmless though and can be submitted to 15.0 - so reassigning to SLE Qt maintainers.