| Summary: | Plasma Desktop Crashes When Adding Weather Widgets (Geolocation) [gps_read, Gpsd::run] | ||
|---|---|---|---|
| Product: | [Unmaintained] plasma4 | Reporter: | Carter Young <ecyoung> |
| Component: | general | Assignee: | Plasma Bugs List <plasma-bugs> |
| Status: | RESOLVED FIXED | ||
| Severity: | crash | CC: | andresbajotierra, andrew5859, bou.gui, bruno, cobexer, damu, dariocambie, EagleScreen, f.leerink, freekdekruijf, jdkaye10, jens, johan.walles, kurt.shingler, m.cole, max7442, mpartap, pswzyu, rb03884, sergeevaas |
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Compiled Sources | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | 4.7.4 | |
| Sentry Crash Report: | |||
| Attachments: |
New crash information added by DrKonqi
New crash information added by DrKonqi New crash information added by DrKonqi New crash information added by DrKonqi New crash information added by DrKonqi New crash information added by DrKonqi New crash information added by DrKonqi |
||
|
Description
Carter Young
2011-07-04 00:16:35 UTC
*** Bug 280545 has been marked as a duplicate of this bug. *** Created attachment 63261 [details]
New crash information added by DrKonqi
plasma-desktop (0.4) on KDE Platform 4.7.41 (4.7.41 (KDE 4.8 >= 20110807) using Qt 4.8.0
- What I was doing when the application crashed:
Trying to start KDE's super plasma shell nested in a xephyr under xfce
- Custom settings of the application:
Maybe it's the public transport widget that i installed and added to the kic... panel bar. Ages ago. When i last could start KDE as a workspace.
-- Backtrace (Reduced):
#6 gps_read (gpsdata=0x1974e10) at libgps_core.c:541
#7 0x00007f45cb3c6e79 in Gpsd::run (this=0x1958ae0) at /usr/src/debug/kde-base/plasma-workspace-9999/plasma-workspace-9999/plasma/generic/dataengines/geolocation/location_gps.cpp:55
#8 0x00007f4642156e87 in QThreadPrivate::start (arg=0x1958ae0) at thread/qthread_unix.cpp:298
[...]
#10 0x00007f4641ec0ebc in start_thread (arg=0x7f45cafa7700) at pthread_create.c:301
#11 0x00007f4640bdad6d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:115
Created attachment 63416 [details]
New crash information added by DrKonqi
plasma-desktop (0.4) on KDE Platform 4.7.1 (4.7.1) using Qt 4.7.3
- What I was doing when the application crashed:
Just adding the weather applet to the panel in a newly generated openSUSE 12.1 M5 system
-- Backtrace (Reduced):
#6 gps_read (gpsdata=0x1dec6d0) at libgps_core.c:541
#7 0x00007f39ff907a03 in Gpsd::run (this=0x1def280) at /usr/src/debug/kde-workspace-4.7.0/plasma/generic/dataengines/geolocation/location_gps.cpp:55
#8 0x00007f3a24a07b25 in QThreadPrivate::start (arg=0x1def280) at thread/qthread_unix.cpp:320
[...]
Created attachment 63723 [details]
New crash information added by DrKonqi
plasma-netbook (0.2) on KDE Platform 4.7.1 (4.7.1) using Qt 4.7.3
- What I was doing when the application crashed:
i tried to start plasma-netbook for the first time, it crashed. I do not have a GPS device.
Bugs #282040 #282039 and #281807 are exact duplicates, is there any information missing that blocks you from fixing this bug?
-- Backtrace (Reduced):
#6 gps_read (gpsdata=0x111dc00) at libgps_core.c:541
#7 0x00007f418e736a03 in Gpsd::run (this=0x10fccd0) at /usr/src/debug/kde-workspace-4.7.0/plasma/generic/dataengines/geolocation/location_gps.cpp:55
#8 0x00007f41b02bbb25 in QThreadPrivate::start (arg=0x10fccd0) at thread/qthread_unix.cpp:320
[...]
[Comment from a bug report cleaner] Adding maintainer of the code to the CC (extracted from https://projects.kde.org/projects/kde/kde-workspace/repository/revisions/master/entry/plasma/generic/dataengines/geolocation/location_gps.cpp) From bug 281807: - What I was doing when the application crashed: I was selecting plasma-netbook workspace from the workspace tab of the kcm, previously selected plasma-desktop worked flawlessly except for the weather desktop disposition (whose backtrace is suggested as releted by dr. konqui) *** Bug 281807 has been marked as a duplicate of this bug. *** *** Bug 282039 has been marked as a duplicate of this bug. *** *** Bug 282040 has been marked as a duplicate of this bug. *** Created attachment 64225 [details]
New crash information added by DrKonqi
plasma-desktop (0.4) on KDE Platform 4.7.1 (4.7.1) using Qt 4.7.4
- What I was doing when the application crashed: Each you try to put weather background plasma crash and restart with the previous background setting.
This setup was made inside a house where you can't get a gps signal.
I don't even know if my computer has a such device. In my case Wifi, bluetooth are all off.
Happen each time.
-- Backtrace (Reduced):
#6 gps_read (gpsdata=0x3043520) at libgps_core.c:541
#7 0x00007fd7c32d1a03 in Gpsd::run (this=0x353d320) at /usr/src/debug/kde-workspace-4.7.1/plasma/generic/dataengines/geolocation/location_gps.cpp:55
#8 0x00007fd7fbb85015 in QThreadPrivate::start (arg=0x353d320) at thread/qthread_unix.cpp:331
[...]
#10 0x00007fd7fb8f4f05 in start_thread (arg=0x7fd7c132d700) at pthread_create.c:301
#11 0x00007fd7fa68f24d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:115
Created attachment 64772 [details]
New crash information added by DrKonqi
plasma-desktop (0.4) on KDE Platform 4.7.2 (4.7.2) using Qt 4.7.4
- What I was doing when the application crashed:
1. Selected Add Widgets
2. Clicked and dragged any weather plasmoid on to the desktop
-- Backtrace (Reduced):
#6 gps_read (gpsdata=0x1f21f90) at libgps_core.c:541
#7 0x00007f306c7c3a03 in Gpsd::run (this=0x1f217e0) at /usr/src/debug/kde-workspace-4.7.2/plasma/generic/dataengines/geolocation/location_gps.cpp:55
#8 0x00007f3095deb015 in QThreadPrivate::start (arg=0x1f217e0) at thread/qthread_unix.cpp:331
[...]
*** Bug 284927 has been marked as a duplicate of this bug. *** Created attachment 64983 [details]
New crash information added by DrKonqi
plasma-desktop (0.4) on KDE Platform 4.7.2 (4.7.2) using Qt 4.7.4
- What I was doing when the application crashed:
- Custom settings of the application:
Adding the Weather widget. The problem appears every time I do that
-- Backtrace (Reduced):
#6 gps_read (gpsdata=0x91be450) at libgps_core.c:541
#7 0xb09b908a in Gpsd::run (this=0x9188df8) at /usr/src/debug/kde-workspace-4.7.2/plasma/generic/dataengines/geolocation/location_gps.cpp:55
#8 0xb6ddb613 in QThreadPrivate::start (arg=0x9188df8) at thread/qthread_unix.cpp:331
[...]
*** Bug 285262 has been marked as a duplicate of this bug. *** *** Bug 285861 has been marked as a duplicate of this bug. *** This crash happens in gpsd, it is very likely it is an upstream bug. I use gpsd-2.95 and I cannot reproduce this crash. *** Bug 285908 has been marked as a duplicate of this bug. *** I can reproduce this crash with gpsd version high than 2.96. I had a review request provided and was already commit to kde-workspace master and 4.7 branch. https://projects.kde.org/projects/kde/kde-workspace/repository/revisions/92b1e7cc6766023050aaaae41390bb5cfcc07ec5 ..how can this even bring down whole plasma-workspace? something is poorly designed here. (In reply to comment #17) > I can reproduce this crash with gpsd version high than 2.96. I had a review > request provided and was already commit to kde-workspace master and 4.7 branch. > > https://projects.kde.org/projects/kde/kde-workspace/repository/revisions/92b1e7cc6766023050aaaae41390bb5cfcc07ec5 Does your patch fix this problem? The commit above should have closed this bug as fixed, but it did not. If your patch really fix this bug then I will close this bug. All plasmoids run on behalf of the plasma-desktop process, if any plasmoid crashes it brings down the entire desktop. That is why we have so many bugs opened for Plasma, although several of them are just duplicates. I was test this on source build with gpsd 2.96, it works fine for me. I had added BUG keyword in my patch then commit, but I don't understand why there didn't closed this bug as well :/ > All plasmoids run on behalf of the plasma-desktop process, if any plasmoid
> crashes it brings down the entire desktop.
Yeah, that's exactly what i referred to as 'poor design'.
Shouldn't the workspace be made a bit more resilient against crashing sub processes? I mean, refusing to *show a desktop environment* _ever again until manual interference_ after a spoilt app has been added is not what one could call 'fail safe' or even 'safe fail' design, is it?
Who's responsible for Plasma architecture again?
(In reply to comment #20) > I was test this on source build with gpsd 2.96, it works fine for me. I had > added BUG keyword in my patch then commit, but I don't understand why there > didn't closed this bug as well :/ You need to have administrator's permissions in bugs.kde.org for the BUG line take effect. I will close this bug then. (In reply to comment #21) > Yeah, that's exactly what i referred to as 'poor design'. > Shouldn't the workspace be made a bit more resilient against crashing sub > processes? I mean, refusing to *show a desktop environment* _ever again until > manual interference_ after a spoilt app has been added is not what one could > call 'fail safe' or even 'safe fail' design, is it? > Who's responsible for Plasma architecture again? Any Plasma developer is aware of this problem. However we cannot change it until KDE 5.0 because we keep binary compatibility throughout a major version (KDE 4 in this case). I do not know exactly who designed plasma-desktop to work this way, but it does not matter, for sure the one who did it is already aware of this problem :-) > Any Plasma developer is aware of this problem. However we cannot change it
> until KDE 5.0 because we keep binary compatibility throughout a major version
> (KDE 4 in this case). I do not know exactly who designed plasma-desktop to work
> this way, but it does not matter, for sure the one who did it is already aware
> of this problem :-)
Well awareness is a good first step indeed...
However, an unknown number of technically inapt users might *right now* sit in front of some shiny wallpaper, waiting for their desktop to appear - in vain. By the time KDE 5.0 will be released, many of them might be deceased already! So maybe someone can come up with a way to deal with it without breaking ABI (for whatever good that is anyways - any product known in the wild that statically link to KDE plasma-desktop?)...
> However we cannot change it [..] because we keep binary compatibility
That by the way is the primary reason windows™ never stopped sucking... still does it C:
*** Bug 288994 has been marked as a duplicate of this bug. *** *** Bug 289161 has been marked as a duplicate of this bug. *** *** Bug 289481 has been marked as a duplicate of this bug. *** *** Bug 289789 has been marked as a duplicate of this bug. *** *** Bug 289771 has been marked as a duplicate of this bug. *** Is there a workaround for this? or do we need to wait for KDE5 to use the KDE desktop again? thanks. (In reply to comment #31) > Is there a workaround for this? or do we need to wait for KDE5 to use the KDE > desktop again? > thanks. The FIXED-IN above says it is fixed in 4.7.4. Have you upgraded to that version and tested the fix? (In reply to comment #32) > (In reply to comment #31) > > Is there a workaround for this? or do we need to wait for KDE5 to use the KDE > > desktop again? > > thanks. > > The FIXED-IN above says it is fixed in 4.7.4. Have you upgraded to that version > and tested the fix? No, I'm using Debian Wheezy/Sid so we are only up to 4.6.5. I would have to start running Experimental in and upgrade a zillion package to get to 4.7.4 running right away. (In reply to comment #33) > No, I'm using Debian Wheezy/Sid so we are only up to 4.6.5. I would have to > start running Experimental in and upgrade a zillion package to get to 4.7.4 > running right away. You can apply yourself the patch mentioned in comment #17 or contact Debian maintainers to do it for you. Honestly, why did you complain that you would have to wait until KDE5 if you do not even bother to read the comments and see that the bug is already fixed for the current KDE SC version? You decided to keep using an old KDE version and not get the fix automatically. *** Bug 289852 has been marked as a duplicate of this bug. *** (In reply to comment #34) > (In reply to comment #33) > > No, I'm using Debian Wheezy/Sid so we are only up to 4.6.5. I would have to > > start running Experimental in and upgrade a zillion package to get to 4.7.4 > > running right away. > > You can apply yourself the patch mentioned in comment #17 or contact Debian > maintainers to do it for you. > > Honestly, why did you complain that you would have to wait until KDE5 if you do > not even bother to read the comments and see that the bug is already fixed for > the current KDE SC version? You decided to keep using an old KDE version and > not get the fix automatically. I'm sorry for annoying you. I read and reread #17 and I don't see any patch. I am not running gpsd now and never have. I do not use bluetooth nor any weather plasmoid. thank you (In reply to comment #36) > I'm sorry for annoying you. I read and reread #17 and I don't see any patch. I > am not running gpsd now and never have. I do not use bluetooth nor any weather > plasmoid. > thank you This bug is about a crash in weather plasmoids. Why did you complain in the first place if you do not use any of them? If you want the patch anyway click on the link in comment #17, then on link "View differences", then on "Also available in: Unified diff". It will download a file called changeset_r92b1e7cc6766023050aaaae41390bb5cfcc07ec5.diff, that is the patch to fix this bug. (In reply to comment #37) > (In reply to comment #36) > > I'm sorry for annoying you. I read and reread #17 and I don't see any patch. I > > am not running gpsd now and never have. I do not use bluetooth nor any weather > > plasmoid. > > thank you > > This bug is about a crash in weather plasmoids. Why did you complain in the > first place if you do not use any of them? Sorry but I didn't move the bug anywhere. I posted the bug and received the number #289789. I believe you moved the bug yourself. If you look at my original bug report, I mentioned nothing about any weather widgets. > > If you want the patch anyway click on the link in comment #17, then on link > "View differences", then on "Also available in: Unified diff". It will download > a file called changeset_r92b1e7cc6766023050aaaae41390bb5cfcc07ec5.diff, that is > the patch to fix this bug. I'll apply the patch and report back. (In reply to comment #38) > Sorry but I didn't move the bug anywhere. I posted the bug and received the > number #289789. I believe you moved the bug yourself. If you look at my > original bug report, I mentioned nothing about any weather widgets. Ok, now I get it. You should have mentioned the original bug before. Anyway, one of your plasmoids is using the geolocation dataengine, which uses the library where this bug happens. (In reply to comment #33) > (In reply to comment #32) > > (In reply to comment #31) > > > Is there a workaround for this? or do we need to wait for KDE5 to use the KDE > > > desktop again? > > > thanks. > > > > The FIXED-IN above says it is fixed in 4.7.4. Have you upgraded to that version > > and tested the fix? > > No, I'm using Debian Wheezy/Sid so we are only up to 4.6.5. I would have to > start running Experimental in and upgrade a zillion package to get to 4.7.4 > running right away. The "proper" fix for this bug has been stated multiple times by the bug maintainer who duplicated your bug. Instead of trying to integrate the patch yourself file a bug on the Debian Bug Tracker page, refer them back to this bug, and as such they will: 1.) Apply the patch into the Debian package, as it is affecting a stable Debian version, and thus justifies aggravating the Debian Maintainers. 2.) Re-prioritize the stabilization of KDE 4.7.4 into the MAIN repository, as this bug is/was rather annoying. Either option will get the ball rolling on Debian's end, as no one there will know until you file a bug related to this one or your original on their bug tracker. (In reply to comment #40) > (In reply to comment #33) > > (In reply to comment #32) > > > (In reply to comment #31) > > > > Is there a workaround for this? or do we need to wait for KDE5 to use the KDE > > > > desktop again? > > > > thanks. > > > > > > The FIXED-IN above says it is fixed in 4.7.4. Have you upgraded to that version > > > and tested the fix? > > > > No, I'm using Debian Wheezy/Sid so we are only up to 4.6.5. I would have to > > start running Experimental in and upgrade a zillion package to get to 4.7.4 > > running right away. > > The "proper" fix for this bug has been stated multiple times by the bug > maintainer who duplicated your bug. Instead of trying to integrate the patch > yourself file a bug on the Debian Bug Tracker page, refer them back to this > bug, and as such they will: > > 1.) Apply the patch into the Debian package, as it is affecting a stable Debian > version, and thus justifies aggravating the Debian Maintainers. > 2.) Re-prioritize the stabilization of KDE 4.7.4 into the MAIN repository, as > this bug is/was rather annoying. > > Either option will get the ball rolling on Debian's end, as no one there will > know until you file a bug related to this one or your original on their bug > tracker. I'm sorry, the "proper way" on any binary based distro, IE Debian, Ubuntu, Fedora(I can't edit my comment). The slowness of response, regarding these types of fixes, by these distros is one of the reasons I switched to a source based distro, like Gentoo or Arch Linux. Thanks Carter. I have filed a bug report to Debian which includes the information contained in my original KDE bug report (#289789). We'll see if this produces anything. In the meantime I'm using fluxbox. *** Bug 291063 has been marked as a duplicate of this bug. *** I'm running KDE 4.74 with openSUSE 12.1 and it is no longer crashing. *** Bug 289847 has been marked as a duplicate of this bug. *** Created attachment 69205 [details]
New crash information added by DrKonqi
plasma-netbook (0.2) on KDE Platform 4.6.5 (4.6.5) using Qt 4.7.4
- What I was doing when the application crashed:
Just turning plasma-desktop to plasma-netbook
-- Backtrace (Reduced):
#6 0x00007fb56c434e8e in gps_sock_read () from /usr/lib/x86_64-linux-gnu/libgps.so.20
#7 0x00007fb56c4313a2 in gps_read () from /usr/lib/x86_64-linux-gnu/libgps.so.20
#8 0x00007fb56c642af3 in Gpsd::run (this=0x15ad1a0) at ../../../../../plasma/generic/dataengines/geolocation/location_gps.cpp:55
#9 0x00007fb59205acf5 in QThreadPrivate::start(void*) () from /usr/lib/libQtCore.so.4
#10 0x00007fb58b5f1b50 in start_thread (arg=<optimized out>) at pthread_create.c:304
|