Bug report #7950
100% reproducible crash when try to move a newly created (point) feature node
|Affected QGIS version:||master||Regression?:||No|
|Operating System:||Easy fix?:||No|
|Pull Request or Patch supplied:||No||Resolution:||fixed|
|Crashes QGIS or corrupts data:||Yes||Copied to github as #:||16814|
That regressed recently. Steps to reproduce:
/!\\ important: you need to check the "[x] Show markers only for selected features" option in the digitizing section of QGIS' options window /!\\
1. Create a new project
2. Insert a new memory points layer (creating new points shapefile also does it, but you can skip the file creation via memory layer)
3. Switch the layer's edit mode on
4. Input two random points on the canvas
5. Switch the layer's edit mode off to save the added points
6. Switch the layer's edit mode back on
7. Try to move one of the two points with the node tool
Over here, on my windows 7 machine, the steps above always lead to QGIS crashing and throwing a minidump.
#1 Updated by Giovanni Manghi almost 10 years ago
- Crashes QGIS or corrupts data changed from No to Yes
#2 Updated by Sandro Santilli almost 10 years ago
Can't reproduce this on GNU/Linux 64bit as of 4f4637913df0cbdaf526d3cccf64f30b9adc4ef3
#3 Updated by Mathieu Pellerin - nIRV almost 10 years ago
So much for a 100% reproducible crasher! ;)
So I've just tried on my linux machine and indeed, linux is crash-free when it comes to this. Will go back to windows and see whether crash occurs on clean profile (i.e. new ~/qgis2)
#4 Updated by Mathieu Pellerin - nIRV almost 10 years ago
Ahh, got it. My windows machine had a not default setting switched on, namely the "[x] Show markers only for selected features" option in the Options' digitizing section. If I switch this off, the crash doesn't occur. If it's switched on, it'll crash on my windows and linux machine. I've updated the above steps to reproduce to mention this.
This option is often given to people reporting slow panning while editing.
#5 Updated by Sandro Santilli almost 10 years ago
My stack trace isn't that good (no line-level debugging symbols) anyway here it is:
Debug: src/app/nodetool/qgsselectedfeature.cpp: 73: (updateGeometry) Entering. Fatal: QGIS died on signal 11 Stacktrace (run through c++filt): output/bin/qgis(_Z15myMessageOutput9QtMsgTypePKc+0x223)[0x549115] /usr/lib/x86_64-linux-gnu/libQtCore.so.4(_Z17qt_message_output9QtMsgTypePKc+0x2e)[0x7f3a9a9d141e] /usr/lib/x86_64-linux-gnu/libQtCore.so.4(+0x718bf)[0x7f3a9a9d18bf] /usr/lib/x86_64-linux-gnu/libQtCore.so.4(_Z6qFatalPKcz+0x94)[0x7f3a9a9d1a64] output/bin/qgis(_Z9qgisCrashi+0x1f)[0x548ef0] /lib/x86_64-linux-gnu/libc.so.6(+0x364a0)[0x7f3a984c14a0] /usr/src/qgis/Quantum-GIS/b/output/lib/libqgis_core.so.1.9.0(_ZN11QgsGeometryC2ERKS_+0x1f)[0x7f3a9bcbdc17] output/bin/qgis(_ZN18QgsSelectedFeature14updateGeometryEP11QgsGeometry+0x114)[0x6e1e88] output/bin/qgis(_ZN18QgsSelectedFeature15createVertexMapEv+0x84)[0x6e4a12] output/bin/qgis(_ZN18QgsSelectedFeature16replaceVertexMapEv+0x24)[0x6e379c] output/bin/qgis(_ZN18QgsSelectedFeature18setSelectedFeatureExP14QgsVectorLayerP12QgsMapCanvas+0x215)[0x6e21cd] output/bin/qgis(_ZN18QgsSelectedFeatureC1ExP14QgsVectorLayerP12QgsMapCanvas+0xe0)[0x6e1a74] output/bin/qgis(_ZN18QgsMapToolNodeTool16canvasPressEventEP11QMouseEvent+0x17d)[0x6de77b] /usr/src/qgis/Quantum-GIS/b/output/lib/libqgis_gui.so.1.9.0(_ZN12QgsMapCanvas15mousePressEventEP11QMouseEvent+0xab)[0x7f3a9b5be7d7] /usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN7QWidget5eventEP6QEvent+0x969)[0x7f3a99eae429] /usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN6QFrame5eventEP6QEvent+0x26)[0x7f3a9a26e3b6] /usr/lib/x86_64-linux-gnu/libQtGui.so.4(_ZN13QGraphicsView13viewportEventEP6QEvent+0x7b)[0x7f3a9a498b2b] /usr/lib/x86_64-linux-gnu/libQtCore.so.4(_ZN23QCoreApplicationPrivate29sendThroughObjectEventFiltersEP7QObjectP6QEvent+0x88)[0x7f3a9aadb028]
#6 Updated by Marco Hugentobler almost 10 years ago
- Assignee set to Marco Hugentobler
I think this bug is related to my recent commit 4c02d7ff9539fce1b4c870d6f9da057f221b7eb0. But I think the mentioned commit is ok, it just reveals a problem that has been there (bug was not causing a crash until now). I'm investigating...
#7 Updated by Marco Hugentobler almost 10 years ago
- Status changed from Open to Resolved
Should be fixed with 6bbc149dfed85c22539a756667e15d9256cafba4. Please test and reopen in case the problem persists.
#8 Updated by Mathieu Pellerin - nIRV almost 10 years ago
- Resolution set to fixed
- Status changed from Resolved to Closed