Andrey

Dear Olaf, just tested version 9.5.0. Looks like 'Coordinate Precision Fix' fixed the broken polygons issue, but this parameter strangely affects the 'Minimum Length' option. In particular, the result includes objects whose length is less than that '-MINLENGTH'. In attachment is the difference between enabled and disabled 'Coordinate Precision Fix' option (black polygons with the option enabled). Also '-PRECISION 1' affects contours if both contours and polygons are generated at the same time, but...

Dear Olaf, I was hoping that the issue was fixed in version 9.4.1, but it is not. This time I used a different dataset and different options. Two tiles from the set had visible problems. The CDM and GUI outputs are still different. Link to a new test-case to reproduce the issue. Hope this helps resolve the issue.

Dear Olaf, I was hoping that the issue was fixed in version 9.4.1, but it is not. This time I used a different dataset and different parameters. Two tiles from the set had visible problems. The CDM and GUI outputs are still different. Link to a new test-case to reproduce the issue. Hope this helps resolve the issue.

Dear Olaf, much improvement has been made in resolving the issue, but it is still not completely fixed. Several tiles from my set still have the broken polygons issue. The CDM and GUI outputs are still different. Tested on version 9.4.0. Link to another test-case. Hope this helps to finally resolve the issue.

But also I would like to note that the number of raster tiles for which the polygon issue occurs has been reduced by almost three times (from half to a fifth).

Hello Olaf. Unfortunately I can't confirm that the issue has been fixed with commit ef34b70b5. For the grid in the test-case provided the fix works, but the issue is not resolved. So I prepared 2 new grids. For 'grid1' the results of processing in CMD and GUI differ in the number of shapes (850 and 849 respectively). And of course the issue with the polygons themselves needs to be fixed. Hope this helps. Link to the new test-case.

Hello Olaf. Unfortunately I can't confirm that the issue has been fixed with commit ef34b70b5. For the grid in the test-case provided the fix works, but the issue is not resolved. So I prepared 2 new grids. For 'grid1' the results of processing in CMD and GUI differ in the number of shapes (850 and 849 respectively). And of course the issue with the polygons themselves needs to be corrected. Hope this helps. Link to the new test-case.

Hi Volker. In my case, this problem occurs for more than half of the raster tiles. You should reproduce the problem with the command saga_cmd shapes_grid 5 -GRID grid.tif -POLYGONS contours_polygon_CMD.shp -INTERVALS 2 -ZLIST "-5000, -4000" and with the same parameters in the GUI in the 'saga-9.3.1_x64.zip' build with this test-case file (also contains my output files).