Open twest820 opened 11 months ago
Can you add a text field (or temporarily reuse one) and use the field calculator to set its value to geom_to_wkt(@geometry)
. What is the value calculated for the offending point?
Oh good point, I should have thought to include that. The WKT's just
'Point (608731.37 4927015.68)'
which is the expected position (which is well within the CRS extents and is surrounded by non-problematic points on all sides). Also as expected, $x
and $y
in the field calculator return that location.
What is the bug or the crash?
Report Details
Python Stack Trace
Stack Trace
Report Details
Python Stack Trace
Stack Trace
System Info CPU Type: x86_64 Kernel Type: winnt Kernel Version: 10.0.19045
Steps to reproduce the issue
These crashes happen when translating or rotating a GeoPackage layer with 2403 points using the move feature and rotate feature tools from the toolbar.
Both crashes are frequent but do not appear to be completely guaranteed. So far if I type in a specific rotation a crash has always followed. Both crashes also occur in QGIS 3.22.14.
In #52877 @jgrocha suggests manually selecting all visible points in the layer and then checking for points which were not selected. In this case 2402 of the points get selected, leaving one point unselected. Unlike #52877, the problem point appears to have valid geometry and the crashes occur regardless of whether or not I put the layer through fix geometries. But, curiously, nothing happens if the point's selected in the attribute table and zoom to feature's chosen from the right click menu. The point's also not displayed next to the points it's actually next to.
I've been able to work around the crashes by recreating the one point and the deleting the original.
Versions
QGIS version 3.28.11-Firenze QGIS code revision b02458a196 Qt version 5.15.3 Python version 3.9.5 GDAL/OGR version 3.7.2 PROJ version 9.3.0 EPSG Registry database version v10.094 (2023-08-08) GEOS version 3.12.0-CAPI-1.18.0 SQLite version 3.41.1 PDAL version 2.5.5 PostgreSQL client version unknown SpatiaLite version 5.1.0 QWT version 6.1.6 QScintilla2 version 2.13.4 OS version Windows 10 Version 2009
Active Python plugins coordinate_capture 0.2 db_manager 0.1.20 grassprovider 2.12.99 MetaSearch 0.3.6 otbprovider 2.12.99 processing 2.12.99 sagaprovider 2.12.99
Supported QGIS version
New profile
Additional context
More user friendly behavior than an AV would be helpful. On the surface it seems clear the digitization tools would benefit from being more conservative in the assumptions they make about layer state—usually the expectation is input error checking should prevent AVs from happening.
However, the same select all but one issue occurs in the original shapefile created in Arc. So my guess here is the chain is more like
I don't have permission to release the data here but if there's something I can do to inspect the problematic point more closely here let me know.