September 2024 - openSUSE KDE 3 - openSUSE Mailing Lists

Mystery Tumbleweed - kmix volume goes to 75% magically sometime after kde3 startup?
by David C. Rankin 20 Oct '24

20 Oct '24

All, Here is another tumbleweed issue I'm trying to track down. The volume with kmix is set to 61%. That's where I set it, and that's where it is set on desktop start. However, after opening konsole, kate, konqueror, firefox, thunderbird, kweather and usually kpatience (patience card game - time killer), at some point, the volume magically jumps to 75% ??? Has anyone else seen there volume magically adjust to 75%? I haven't isolated which app does it. Whenever it happens, I don't know about it until the next sound is played -- usually the konsole error telling me I hit backspace one too many times. Thought I would ask... -- David C. Rankin, J.D.,P.E.

2 7

Test build of kdelibs3 with rust alert patch - I'll test
by David C. Rankin 17 Oct '24

17 Oct '24

Yasuhiko, Since we have eliminated the openssh and other kdelibs3/qt3 issues, can you make a private build of kdelibs3 with the rust.xml alert patch and post a link for me to test to see if the Embedded Advanced Text Editor katepart issue has resolved? I'm curious about what the issue was there. The change to rust.xml is innocuous and should make no difference whatsoever to the katepart. Not a big issue, but I would like to see the rust.xml syntax file updated. Not that I'm a big proponent of the rust craze, but when looking at rust sources those highlighted comments help readability. Whenever you have time, just let me know and I'm happy to test. -- David C. Rankin, J.D.,P.E.

2 4

.xsession-errors UPower? Anybody else get this with openSUSE Firefox build?
by David C. Rankin 29 Sep '24

29 Sep '24

All, After the last sound crash (uptime 2 days 20 hours), checking .xsession-errors I ran across the following warning (not sound related): [Parent 10088, Main Thread] WARNING: Failed to enumerate devices of org.freedesktop.UPower: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name is not activatable : 'glib warning', file /home/abuild/rpmbuild/BUILD/firefox-130.0.1/toolkit/xre/nsSigHandlers.cpp:187 ** (firefox:10088): WARNING **: 01:48:08.287: Failed to enumerate devices of org.freedesktop.UPower: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name is not activatable It's firefox and not kde3, but is there some new signal handler we need to add? Or should we just ignore this warning? This is firefox from OSS. -- David C. Rankin, J.D.,P.E.

3 14

NEW BUG - sftp kio is BROKEN
by David C. Rankin 23 Sep '24

23 Sep '24

Yasuhiko, I went to use konqueror to browse another Linux box on my LAN as I've done 1000's of times, ctrl + t for a new tab and then: sftp://valkyrie/home/david sftp:// is broken. Usually as you are typing, sftp is so good it autocompletes your path, e.g. sftp://valkyrie/home/d would shown 'david' for completion and you can simply select it and complete the connection. Now, sftp is just broken. See the following link where I attempted to open a saved kate session that opens files from the host on my LAN: https://paste.opensuse.org/pastes/d66c92f7bb18 This is a "hair on fire" moment. We have to figure out what broke and when. It's probably been a couple of weeks since I worked on the remote project, so I don't know what day it broke, but something broke (maybe the new openssh?) The journal shows a coredump: Sep 22 03:36:44 wizard systemd-coredump[13644]: [🡕] Process 13533 (kio_sftp) of user 1000 dumped core. Stack trace of thread 13533: #0 0x00007fcfe00826a3 _ZN11QStringListD1Ev (kio_sftp.so + 0xa6a3) #1 0x00007fcfdfa43c16 __run_exit_handlers (libc.so.6 + 0x43c16) #2 0x00007fcfdfa43d60 __GI_exit (libc.so.6 + 0x43d60) #3 0x00007fcfe1d787cb _ZN3KIO9SlaveBase15sigsegv_handlerEi (libkio.so.4 + 0x1787cb) #4 0x00007fcfdfa41290 __restore_rt (libc.so.6 + 0x41290) #5 0x00007fcfe0fd5dac _ZN7QStringaSERKS_ (libqt-mt.so.3 + 0x5d5dac) #6 0x00007fcfe008f068 _ZN11KSshProcess7getLineEv (kio_sftp.so + 0x17068) #7 0x00007fcfe008f9fd _ZN11KSshProcess7connectEv (kio_sftp.so + 0x179fd) #8 0x00007fcfe009086e _ZN12sftpProtocol14openConnectionEv.part.0 (kio_sftp.so + 0x1886e) #9 0x00007fcfe008ac45 _ZN12sftpProtocol3getERK4KURL (kio_sftp.so + 0x12c45) #10 0x00007fcfe1dae478 _ZN3KIO9SlaveBase8dispatchEiRK9QMemArrayIcE (libkio.so.4 + 0x1ae478) #11 0x00007fcfe1dac371 _ZN3KIO9SlaveBase12dispatchLoopEv (libkio.so.4 + 0x1ac371) #12 0x00007fcfe0085489 kdemain (kio_sftp.so + 0xd489) #13 0x000055d01bf90d7d _ZL6launchiPKcS0_S0_iS0_bS0_bS0_ (kdeinit + 0x8d7d) #14 0x000055d01bf91f60 _ZL23handle_launcher_requesti (kdeinit + 0x9f60) #15 0x000055d01bf9254a _ZL15handle_requestsi (kdeinit + 0xa54a) #16 0x000055d01bf8e9fb main (kdeinit + 0x69fb) #17 0x00007fcfdfa2a340 __libc_start_call_main (libc.so.6 + 0x2a340) #18 0x00007fcfdfa2a409 __libc_start_main_impl (libc.so.6 + 0x2a409) #19 0x000055d01bf8f185 _start (kdeinit + 0x7185) ELF object binary architecture: AMD x86-64 Sep 22 03:36:44 wizard systemd[1]: systemd-coredump(a)5-13638-0.service: Deactivated successfully. Sep 22 03:36:44 wizard systemd-coredump[13641]: [🡕] Process 13517 (kio_sftp) of user 1000 dumped core. Stack trace of thread 13517: #0 0x00007fcfe00826a3 _ZN11QStringListD1Ev (kio_sftp.so + 0xa6a3) #1 0x00007fcfdfa43c16 __run_exit_handlers (libc.so.6 + 0x43c16) #2 0x00007fcfdfa43d60 __GI_exit (libc.so.6 + 0x43d60) #3 0x00007fcfe1d787cb _ZN3KIO9SlaveBase15sigsegv_handlerEi (libkio.so.4 + 0x1787cb) #4 0x00007fcfdfa41290 __restore_rt (libc.so.6 + 0x41290) #5 0x00007fcfe0fd5dac _ZN7QStringaSERKS_ (libqt-mt.so.3 + 0x5d5dac) #6 0x00007fcfe008f068 _ZN11KSshProcess7getLineEv (kio_sftp.so + 0x17068) #7 0x00007fcfe008f9fd _ZN11KSshProcess7connectEv (kio_sftp.so + 0x179fd) #8 0x00007fcfe009086e _ZN12sftpProtocol14openConnectionEv.part.0 (kio_sftp.so + 0x1886e) #9 0x00007fcfe008ac45 _ZN12sftpProtocol3getERK4KURL (kio_sftp.so + 0x12c45) #10 0x00007fcfe1dae478 _ZN3KIO9SlaveBase8dispatchEiRK9QMemArrayIcE (libkio.so.4 + 0x1ae478) #11 0x00007fcfe1dac371 _ZN3KIO9SlaveBase12dispatchLoopEv (libkio.so.4 + 0x1ac371) #12 0x00007fcfe0085489 kdemain (kio_sftp.so + 0xd489) #13 0x000055d01bf90d7d _ZL6launchiPKcS0_S0_iS0_bS0_bS0_ (kdeinit + 0x8d7d) #14 0x000055d01bf91f60 _ZL23handle_launcher_requesti (kdeinit + 0x9f60) #15 0x000055d01bf9254a _ZL15handle_requestsi (kdeinit + 0xa54a) #16 0x000055d01bf8e9fb main (kdeinit + 0x69fb) #17 0x00007fcfdfa2a340 __libc_start_call_main (libc.so.6 + 0x2a340) #18 0x00007fcfdfa2a409 __libc_start_main_impl (libc.so.6 + 0x2a409) #19 0x000055d01bf8f185 _start (kdeinit + 0x7185) ELF object binary architecture: AMD x86-64 Any ideas? Or do you want a bug report opened? Let me know. This may just need to be option changes to use the new openssh? -- David C. Rankin, J.D.,P.E.

2 6

KDE3 Konqueror FileManagement Loses File Associations Randomly
by David C. Rankin 19 Sep '24

19 Sep '24

All, Yasuhiko, On occasions konqueror filemanagement loses its file associations to various type files randomly. Here is an example of where it doesn't know what an .xcf image opens with despite listing the file type as a "GIMP Native Image..." https://paste.opensuse.org/pastes/4a65237dfc92 This confusion is on a per-session basis and will be reset to normal on next logout/login. This looks suspiciously like a memory corruption error of some type, a use-after-free, etc... It obviously initializes the file-associations correctly on startup, but then some time (could be hours, weeks, or never) this corruption pops up. This time it is also occurring after the random "Knotify Crash" that disables the local sound theme I reported earlier. That crash may be doing more than just disabling the local sound theme and preventing sox from playing sound files (aplay continues to work). The crash most often occurs when pressing Backspace in konsole when already at the beginning of the line (triggering the system beep). That seems to be what actually causes the Knotify crash. (maybe the beep and the sound-theme error try and use the sound system at the same time causing the crash or the use of a resource right after it has been freed?) What debug packages do I need to load to actually get a full backtrace of the crash? Before Tumbleweed, I would at least get a backtrace in kcrash. In Tumbleweed it just tells me it can't get a backtrace because the packages are built in a way that doesn't provide them (I take that as I'm missing the .debug package needed) If anybody else has answers, let me know. If anybody else triggers the Knotify crash - take note of what was happening at the time and let us know. One of these years we will find the cause :) -- David C. Rankin, J.D.,P.E.

2 17

[kde3] kdmrc logging misrepresentations
by Felix Miata 15 Sep '24

15 Sep '24

I've been seeing the following in logs since forever, or at least, since I created a symlink where admin config overrides belong, in /etc/: # journalctl -b --no-host | grep ultipl Aug 29 23:35:18 kdm_config[895]: Multiple occurrences of key 'ShowUsers' in section [X-*-Greeter] of /opt/kde3/share/config/kdm/kdmrc Aug 29 23:35:18 kdm_config[895]: Multiple occurrences of key 'UseTheme' in section [X-*-Greeter] of /opt/kde3/share/config/kdm/kdmrc Aug 29 23:35:18 kdm_config[895]: Multiple occurrences of key 'AllowNullPasswd' in section [X-:*-Core] of /opt/kde3/share/config/kdm/kdmrc Aug 29 23:35:18 kdm_config[895]: Multiple occurrences of key 'AllowShutdown' in section [X-:*-Core] of /opt/kde3/share/config/kdm/kdmrc # Might there be a way to make this not happen without eliminating /my/ kdmrc presence in /etc/? # ls -gG /etc/kde/config/kdm lrwxrwxrwx 1 34 Jul 23 2023 /etc/kde/config/kdm -> ../../../opt/kde3/share/config/kdm # egrep '^ShowUsers|^UseTheme|^AllowNullPasswd|^AllowShutdown' /etc/kde/config/kdm/kdmrc ShowUsers=Selected UseTheme=false AllowNullPasswd=true AllowShutdown=All # -- Evolution as taught in public schools is, like religion, based on faith, not based on science. Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata

2 1

Add Alerts (FIXME, TODO, etc..) to rust katepart/syntax/rust.xml
by David C. Rankin 12 Sep '24

12 Sep '24

Yasuhiko, The current rust.xml katepart syntax file does not highlight alerts like TODO, FIXME, etc. in the comments. The katepart/syntax file needs to be updated to include the katepart/syntax/alerts.xml file. A patch to do so is: diff -ruNb a/syntax/rust.xml b/syntax/rust.xml --- a/syntax/rust.xml 2015-10-28 15:38:53.000000000 -0500 +++ b/syntax/rust.xml 2024-09-10 08:53:59.108745900 -0500 @@ -323,10 +323,14 @@ <RegExpr String="U[0-9a-fA-F]{8}" attribute="CharEscape" context="#pop"/> <RegExpr String="." attribute="Error" context="#pop"/> </context> - <context attribute="Comment" lineEndContext="#pop" name="Commentar 1"/> + <context attribute="Comment" lineEndContext="#pop" name="Commentar 1"> + <LineContinue attribute="Comment" context="#stay"/> + <IncludeRules context="##Alerts" /> + </context> <context attribute="Comment" lineEndContext="#stay" name="Commentar 2"> <DetectSpaces/> <Detect2Chars char="*" char1="/" attribute="Comment" context="#pop" endRegion="Comment"/> + <IncludeRules context="##Alerts" /> </context> </contexts> <itemDatas> @@ -348,6 +352,7 @@ <itemData name="Macro" defStyleNum="dsOthers"/> <itemData name="Attribute" defStyleNum="dsOthers"/> <itemData name="Lifetime" defStyleNum="dsOthers" bold="1"/> + <itemData name="Alert" defStyleNum="dsAlert" spellChecking="false"/> <itemData name="Error" defStyleNum="dsError"/> </itemDatas> </highlighting> Can that be applied to the system wide rust.xml in /opt/kde3/share/apps/katepart/syntax/rust.xml so that it is available? -- David C. Rankin, J.D.,P.E.

2 3

BUG - focus-follows-mouse broken for Gtk apps in latest builds of KDE3
by David C. Rankin 12 Sep '24

12 Sep '24

All, I just did a fresh install of Tumbleweed. I have both 15.4 and Tumbleweed on separate disks I can boot back and forth between. My 15.4 KDE3 hasn't been updated in several months -- it continues to work perfectly. However, whatever has changed (base libraries, etc..) in the KDE3 build has broken the focus-follows-mouse and middle-mouse-scroll over Gtk apps such as Firefox/Thunderbird unless those windows are raised to the top level. (in other words the mouse-wheel will not scroll the contents of FF or TB when other windows are above them) I've done quite a bit of digging and filed a bug with mozilla -- though it should properly be an openSUSE KDE3 bug is what it looks like. The bug I filed is: Firefox Bug - no middle-mouse scroll on focused window not raised to top https://bugzilla.mozilla.org/show_bug.cgi?id=1907708 See also discussion in the opensuse-user list thread: "Why does mozilla repo have FF 115.12.0 for 15.4 and 115.11.0 for TW?" This is not a new issue and has been around for years in various forms, so this may be a regression in a library or environment variable that feed into the KDE3 build. The earlier reports of the same problem: Scrolling inactive windows in KDE/KWin with mouse wheel https://bbs.archlinux.org/viewtopic.php?id=239918 Mousewheel action over Firefox doesn't work if Firefox window is not active https://bugzilla.mozilla.org/show_bug.cgi?id=1359226 A work-around is to pass GDK_CORE_DEVICE_EVENTS=1 when starting firefox, e.g. $ GDK_CORE_DEVICE_EVENTS=1 firefox-esr & Focus-follows-mouse and mouse-wheel-scroll then work normally in FF and TB on KDE3. (all other non-gtk apps didn't have this issue to begin with). Moreover, this is KDE3 unique. If I launch fluxbox as the desktop on TW, the focus-follows-mouse and mouse-wheel-scroll work normally in FF and TB without setting GDK_CORE_DEVICE_EVENTS, so this is a KDE3 bug and it is recent. KDE3 on 15.4 does not have this problem. The base packages there are from Sep 5, 2023, e.g. -rw-r--r-- 1 root root 15656560 Sep 5 2023 kdelibs3-3.5.10-lp154.264.1.x86_64.rpm -rw-r--r-- 1 root root 1097892 Sep 5 2023 kdebase3-runtime-3.5.10.1-lp154.379.1.x86_64.rpm -rw-r--r-- 1 root root 9781396 Sep 5 2023 kdebase3-3.5.10.1-lp154.379.1.x86_64.rpm So what is the consensus, should I open a bug report on openSUSE for this or can it be investigated and fixed from here? Let me know. -- David C. Rankin, J.D.,P.E.

1 1

TW - ksysguard broken - not updating - please confirm
by David C. Rankin 10 Sep '24

10 Sep '24

Yasuhiko, All, I just had the relatively normal "Knotify Crash" and was going to look at ksysguard to see if there was anything funny going on. Starting ksysguard, the window opens and all looks fine -- but nothing is updating, it's just dead. Can somebody open ksysguard on Tumbleweed and see if it updates on your system. My normal "Knotify Crash" is the old random crash that disables system sounds for your sound theme: jack_client_new: deprecated Cannot connect to server socket err = No such file or directory Cannot connect to server request channel jack server is not running or cannot be started JackShmReadWritePtr::~JackShmReadWritePtr - Init not done for -1, skipping unlock JackShmReadWritePtr::~JackShmReadWritePtr - Init not done for -1, skipping unlock akode: Guessed format: xiph KNotify::playTimeout KNotify::playTimeout KCrash: Application 'knotify' crashing... X Error: BadWindow (invalid Window parameter) 3 Major opcode: 19 Minor opcode: 0 Resource id: 0xc0000b X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d021bd X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d021bd X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d025f1 X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d025f1 libpng warning: iCCP: known incorrect sRGB profile X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d02bcb X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d02bcb X Error: BadWindow (invalid Window parameter) 3 Major opcode: 20 Minor opcode: 0 Resource id: 0x1d02bcb X Error: BadWindow (invalid Window parameter) 3 Major opcode: 19 Minor opcode: 0 Resource id: 0x1800630 It may run a week without crashing, or it may crash within a couple of hours of booting. Just random. I don't know if the knotify crash effects ksysguard updating, but I thought I would include it for completeness. If you have other ideas what to check let me know. ps still shows knotify running, but the sound theme is dead, e.g. 13:38 wizard:~> ps axf | grep notify 89933 pts/2 S+ 0:00 | \_ grep --color=auto notify 83171 ? S 0:00 knotify [kdeinit] Also, does anybody know how to restart the sound theme or sound server from the command line? I'd rather do that than logging out/in if possible. Thanks! -- David C. Rankin, J.D.,P.E.

1 0

YaST - KDE 3 Desktop Environment - Pattern Broken by kde3-gtk-qt-engine
by David C. Rankin 05 Sep '24

05 Sep '24

All, Yasuhiko, Responding to a question by Felix regarding resume= for suspend, he asked for the patterns installed on my system. Checking yast I stumbled across the actual pattern KDE 3 Desktop Environment. (I'd never looked, I just presumed it was still missing as it had been for years) It's there! That's good! The bad is you cannot install it due to libbonoboui. Attempting to select the pattern results in the dependency error: ! nothing provides 'libbonoboui-2.so.0()(64bit)' needed by the to be installed kde3-gtk-qt-engine-0.8svn.20071009-52.33.x86_64 I know I've flagged the missing dep before, but there has to be some way to bring that library into the kde3 repo or patch the kde3-gtk-qt-engine to build with the newer version because that is badly needed for the kde3 continuity of desktop look and feel across gtk apps. Yes many gtk apps are moving to gtk4 being ported from gtk+3 and/or gtk+2, but many (like Gimp) still use gtk+2/gtk+3. (gtk+3 removed the ruler widget Gimp relies on, so Gimp will continue to need gtk+2 for some time) We can probably remove the following packages from the pattern as well unless the underlying libraries have been updated (the packages stay for install if individually chosen, but the likely don't need to be in the Pattern): kdebase3-extra kdenetwork3 kdenetwork3-IRC (most of the services don't exist anymore) kdenetwork3-lan kdenetwork3-lisa The rest is fine, kde3-knemo works quite nicely still. Not a high priority -- but the Pattern should install if chosen from YaST, and kde3-kcm_gtk isn't an viable alternative for kde3-gtk-qt-engine. -- David C. Rankin, J.D.,P.E.

3 3