Bug#855486: libkf5purpose-bin: inappropriate install dependency on nodejs for some architectures
Package: libkf5purpose-bin Version: 1.1-4 Severity: important Dear Maintainer, For several architectures, including at least alpha and hppa, there is no "nodejs" package available. For those architectures, "libkf5purpose-bin" has an installation dependency on "nodejs" that needs to be removed. The above action has supposedly already been taken for "armel", which also does not have "nodejs" built. Respectfully, --Bob -- System Information: Debian Release: 9.0 APT prefers unreleased APT policy: (500, 'unreleased'), (500, 'unstable') Architecture: alpha Kernel: Linux 4.9.0 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash Init: systemd (via /run/systemd/system)
Bug#648994: qt4-x11: g++-4.4 optimizer broken on Alpha
Source: qt4-x11 Version: 4:4.7.3-8 Severity: important Tags: patch Dear Maintainer, Due to a bug in the g++-4.4 optimizer, the normal "-O2" build of "qt4-x11" binary packages results in segfaults on the alpha architecture when apps are run that use the "qt4" libraries. This includes "kdm" and the KDE4 "plasma-desktop" at a minimum (bad behavior verified and easy to reproduce). A verified workaround is to build with "-O0". Note that DEB_BUILD_OPTIONS=noopt is ignored by the QT build system: I ended up editing the appropriate files under "mkspecs" directly to test the workaround. Unknown whether use of gcc/g++-4.6 fixes the issue. -- System Information: Debian Release: wheezy/sid APT prefers unreleased APT policy: (500, 'unreleased'), (500, 'unstable') Architecture: alpha Kernel: Linux 3.2.0-rc1 Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/2016180645.9530.54794.report...@smirkin.frus.com
KDE4 testing on Alpha
Armed with nearly three times the RAM I used to have, I screwed up my courage and tried a plasma-desktop session on the console. It locked up the machine as it did on all prior attempts, but now that the machine isn't paging itself to death, it seems easier to start narrowing down cause and effect. In particular, the desktop session was running great... Had previously- saved terminal window sessions open and was keeping track of physical memory and swap usage. Just under 100% of 1.5 GB RAM was in use, although much of it was buffering: only a bit over 8 KB of swap was in use. Then, "virtuoso-t" started executing the background. The syslog up to the point of having to press the reset switch is showing tens of thousands of suppressed callbacks, with log entries spaced roughly six seconds apart. "tracker-store" similarly generated lots of suppressed callback messages, but only numbered in the thousands rather than tens of thousands. All of these symptoms are unique to console plasma-desktop sessions. FreeNX plasma-desktop sessions run fine and don't cause the machine any grief that I can tell. An obvious difference between the two session types is the display, and the fact that kernel modesetting is enabled for the console. Kernel is currently the standard kernel.org 3.2.0-rc4. Most of the KDE packages are version 4:4.6.5-n: I do "apt-get upgrade" daily and am running "unstable/sid". The console display is a radeon 7500 (RV200 QW). I'd be happy to open a bug report if someone could provide guidance as to an appropriate package. For now, I'm copying this to the "debian-qt-kde" list in the hope of attracting the right kind of attention :-). Thanks in advance for any help in getting this issue solved. --Bob -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20111209150209.ga1...@gherkin.frus.com
Re: KDE4 testing on Alpha
On Sat, Dec 10, 2011 at 08:03:46PM +0100, Pino Toscano wrote: > If it's the virtuoso process to create issue, you can try disabling the > semantic desktop (which uses the virtuoso DB): either you uninstall the > virtuoso packages, or disable nepomuk for your user in a console with: > $ kwriteconfig --file nepomukserverrc --group "Basic Settings" \ > --key "Start Nepomuk" --type bool false > (you can also do that in the "semantic desktop" section of > `systemsettings`) > > If that does not avoid the hard lockup, I don't have more ideas at the > moment... Thank you for your reply. Disabling nepomuk did not seem to have any effect. In fact, with the 3.2-rc5 kernel, the lock-up happened even more quickly. When starting KDE4, there's the splash screen and five icons that fade-in from left to right. At the point where the big "K" icon finishes fading-in, that's where the system locked up this time. It has done so at this precise point in the past, which is why I suspect issues with KMS/drm. For what it's worth, Gnome-3 (which requires KMS) isn't working either: I can't test Gnome-3 components using FreeNX because there's no KMS for that type of display, so I get "fallback mode" (Gnome-2) by default. I'll go ahead and reenable nepomuk, and then try another KDE4 FreeNX session. Would it be correct to assume nepomuk/virtuoso isn't the issue if I see those processes running for the FreeNX desktop session? Trying to debug large packages where there may be multiple failure modes in play is difficult. I'm trying to determine all the differences between console and non-console (e.g., FreeNX) GUI sessions, but the most obvious difference is kernel modesetting. KDE session differences due to whether KMS is available/enabled would be nice to know about, but my level of knowledge as to KDE internals is limited at best :-(. --Bob -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20111211135852.ga31...@gherkin.frus.com
Re: KDE4 testing on Alpha
On Mon, Dec 12, 2011 at 09:34:05PM +1300, Michael Cree wrote: > Bob, I've just recompiled iceweasel and icedove with gcc-4.6 and full > optimisation (other than --no-relax in the linker) and the crashes in > those two packages have disappeared. I am sending this message with > icedove from my Alpha --- the first time I have been able to do this > since before the release of Lenny! I suspect that optimisation bugs in > gcc-4.4 have been fixed in gcc-4.5 and gcc-4.6 but not back-ported to > gcc-4.4, and on Alpha we are unfortunately still defaulting to gcc-4.4. > I wonder if the crashes in KDE are due to bugs in gcc-4.4. Wonderful news concerning gcc/g++-4.6, and I think rebuilding several of our (Alpha's) more "questionable" packages would be a worthwhile exercise. That being said, my FreeNX KDE4 sessions are fully operational (with non-optimized qt4-x11): no system lock-ups, nepomuk running, virtuoso-t running, etc. Got little pop-up status messages about indexing the desktop starting, then completing approximately a minute and a half later without any problems. There may be other bugs lurking within the KDE packages, but the system is rock-solid stable as long as I don't try KDE4 or Gnome-3 on the console. None of this means we aren't seeing some buggy kernel behavior (kms/drm) due to kernels built with gcc-4.4 :-(. I think trying a kernel built with gcc-4.6 would be a worthwhile exercise, so I'll do that with the next "-rcX" when that happens. Have I mentioned lately how much I hate an untrustworthy tool chain? Apologies to the kde/qt maintainers for the noise. --Bob -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20111212125823.ga4...@gherkin.frus.com
Re: KDE4 testing on Alpha
On Wed, Dec 14, 2011 at 08:02:32PM +1300, Michael Cree wrote: > On 13/12/11 01:58, Bob Tracy wrote: > > Wonderful news concerning gcc/g++-4.6, and I think rebuilding several of > > our (Alpha's) more "questionable" packages would be a worthwhile > > exercise. > > We are now defaulting to gcc-4.6 on Alpha. I have binNMU qt4-x11 to > rebuild with gcc-4.6. Are there any other KDE packages that should be > binNMU in the hope that we can rid the KDE crashes? Rebuilding that one source package was enough to take care of all the KDE4 crashes I've seen *other* than the console-based sessions, which are probably KMS-related. With the qt4-x11 packages built using the 4.4 compiler *with* optimization, even "kdm" would segfault. Let me know when the new qt4-x11 binary packages are ready, and I'll release the holds on my current custom-built packages (to give the new ones a try). Thanks! --Bob -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20111214073506.ga9...@gherkin.frus.com
Re: KDE4 testing on Alpha
On Wed, Dec 14, 2011 at 11:04:11PM +0100, Pino Toscano wrote: > > (KDE4 causing hard system lockups/freezes during console > > sessions, but not FreeNX sessions.) > > Hm I see. > My next hint would be try to disable the compositing of kwin (the window > manager), maybe it somehow detects it is available and crashes when > being used or crashes even when trying to detect it; > $ kwriteconfig --file kwinrc --group Compositing --key Enabled \ > --type bool false That worked! For the first time in more months than I can remember, I'm actually running at least a minimal version of one of the major graphical environments on my Alpha console. At the risk of being perceived as greedy :-), what can I do to help as far as getting compositing working? --Bob -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20111220124438.ga4...@gherkin.frus.com
Bug#541402: alpha: plasma segfault on kde4 startup
Package: kdebase-workspace-bin Version: 4:4.2.4-1 Severity: grave Justification: renders package unusable Since the upgrade from KDE3 to KDE4 on the Alpha platform, KDE has been unusable. "/usr/bin/plasma" consistently segfaults on startup. This may or may not be related to #519518: the reporter was using an i686 platform. The following diagnostic output was captured (was not given the opportunity to save it directly -- "save" button grayed-out): This backtrace appears to be of no use. This is probably because your packages are built in a way which prevents creation of proper backtraces, or the stack frame was seriously corrupted in the crash. (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) (no debugging symbols found) [Thread debugging using libthread_db enabled] [New Thread 0x20004bad580 (LWP 9690)] 0x021bd638 in nanosleep () from /lib/libc.so.6.1 [Current thread is 0 (process 9690)] Thread 1 (Thread 0x20004bad580 (LWP 9690)): #0 0x021bd638 in nanosleep () from /lib/libc.so.6.1 #1 0x021bd428 in sleep () from /lib/libc.so.6.1 #2 0x02bbef60 in ?? () from /usr/lib/libkdeui.so.5 #3 0x02bbfedc in KCrash::defaultCrashHandler () from /usr/lib/libkdeui.so.5 #4 0x02140db0 in ?? () from /lib/libc.so.6.1 #5 0x0200020c6e80 in ?? () from /usr/lib/libQtCore.so.4 #6 0x0200020c8508 in QObject::QObject () from /usr/lib/libQtCore.so.4 #7 0x020001316cb8 in QAction::QAction () from /usr/lib/libQtGui.so.4 #8 0x0200013a916c in QWidgetAction::QWidgetAction () from /usr/lib/libQtGui.so.4 #9 0x020001ba33f8 in ?? () from /usr/lib/libQtGui.so.4 #10 0x020001ba61a8 in ?? () from /usr/lib/libQtGui.so.4 #11 0x020001ba6de4 in ?? () from /usr/lib/libQtGui.so.4 #12 0x020001b90898 in QGraphicsLinearLayout::sizeHint () from /usr/lib/libQtGui.so.4 #13 0x020001b8e88c in ?? () from /usr/lib/libQtGui.so.4 #14 0x020001b8ef80 in QGraphicsLayoutItem::effectiveSizeHint () from /usr/lib/libQtGui.so.4 #15 0x020001b940f8 in QGraphicsWidget::sizeHint () from /usr/lib/libQtGui.so.4 #16 0x020001b8e88c in ?? () from /usr/lib/libQtGui.so.4 #17 0x020001b8ef80 in QGraphicsLayoutItem::effectiveSizeHint () from /usr/lib/libQtGui.so.4 #18 0x020001b96bc0 in QGraphicsWidget::setGeometry () from /usr/lib/libQtGui.so.4 #19 0x020015cf16c8 in ?? () from /usr/lib/kde4/plasma_applet_systemtray.so #20 0x020001b939f4 in QGraphicsWidget::resize () from /usr/lib/libQtGui.so.4 #21 0x020001b94690 in QGraphicsWidget::updateGeometry () from /usr/lib/libQtGui.so.4 #22 0x020001b8fc24 in QGraphicsLayoutItem::setMinimumSize () from /usr/lib/libQtGui.so.4 #23 0x0233f37c in Plasma::Applet::setBackgroundHints () from /usr/lib/libplasma.so.3 #24 0x020015cf41c8 in ?? () from /usr/lib/kde4/plasma_applet_systemtray.so #25 0x020015cf56cc in KPluginFactory::createInstance () from /usr/lib/kde4/plasma_applet_systemtray.so #26 0x02000109eab0 in KPluginFactory::create () from /usr/lib/libkdecore.so.5 #27 0x0234a8c8 in Plasma::Applet::load () from /usr/lib/libplasma.so.3 #28 0x0208d5e0 in ?? () from /usr/lib/libkdeinit4_plasma.so #29 0x0208fb38 in ?? () from /usr/lib/libkdeinit4_plasma.so #30 0x0236f110 in Plasma::Corona::initializeLayout () from /usr/lib/libplasma.so.3 #31 0x020b141c in ?? () from /usr/lib/libkdeinit4_plasma.so #32 0x020b1e00 in ?? () from /usr/lib/libkdeinit4_plasma.so #33 0x020b6e78 in ?? () from /usr/lib/libkdeinit4_plasma.so #34 0x0200020d0d64 in QMetaObject::activate () from /usr/lib/libQtCore.so.4 #35 0x0200020d1f78 in QMetaObject::activate () from /usr/lib/libQtCore.so.4 #36 0x0200020d9db8 in ?? () from /usr/lib/libQtCore.so.4 #37 0x0200020d9f30 in ?? () from /usr/lib/libQtCore.so.4 #38 0x0200020c925c in QObject::event () from /usr/lib/libQtCore.so.4 #39 0x02000131f574 in QApplicationPriv
Bug#541402: alpha: plasma segfault on kde4 startup
ACK on the below indicated packages. I won't have any free time to devote to this until this coming Thursday, but I *will* do what I can to help. Thanks! --Bob On Mon, Aug 31, 2009 at 12:55:35AM +0200, Sune Vuorela wrote: > tag 541402 +help > thanks > > On Thursday 13 August 2009 23:24:38 Bob Tracy wrote: > > Package: kdebase-workspace-bin > > Version: 4:4.2.4-1 > > Severity: grave > > Justification: renders package unusable > > Hi > > Currently, the debian-qt-kde team is out of alpha machines, so we kind of > need > help from the alpha people. > > > This backtrace appears to be of no use. > > And it is. > > PLease install at least kdebase-workspace-bin and kde4libs-dbg and try again. > > We do probably need someone of the alpha people to do some actual debugging > here. > > /Sune -- To UNSUBSCRIBE, email to debian-qt-kde-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
