[QGIS-Developer] Did polygon winding order changed in QGIS or GDAL recently?

[Richard Duivenvoorde]

Recently there was a question on the user list, about the winding order of polygons in json export, which seemed/(was thought to have) changed in time:

https://lists.osgeo.org/pipermail/qgis-user/2024-August/054721.html

And now the open data vector tiles of the Netherlands, do not work anymore in QGIS (in dutch):
https://geoforum.nl/t/bgt-vectortiles-in-qgis-zijn-niet-meer-compleet/9731

Their conclusion is: in the standards the OGC winding order is different from the MVT winding order, that is why QGIS (thinking QGIS choose for OGC), is not rendering the MVT tiles anymore...

So my question is:

- anybody aware of changes of the winding order?
- or do we do things 'better' now, breaking older functionality?
- or are the pdok mvt tiles wrongly ordered?

In the case of MVT tiles (haven't looked into the specs though), but it seems that the reader of the data would respect ordering differences between standards?

If you want to checkout the pdok/dutch open data:

https://api.pdok.nl/lv/bgt/ogc/v1/tiles/WebMercatorQuad/{z}/{y}/{x}?f=mvt
CRS: EPSG:3857
Max Zoom Level: 17
Styling:
https://api.pdok.nl/lv/bgt/ogc/v1/styles/bgt_achtergrondvisualisatie__webmercatorquad?f=mapbox

In favour of QGIS, an older Esri osm dataset, is still visible/working:

https://basemaps.arcgis.com/arcgis/rest/services/OpenStreetMap_v2/VectorTileServer
Max Zoom Level: 19
https://www.arcgis.com/sharing/rest/content/items/3e1a00aeae81496587988075fe529f71/resources/styles/root.json?f=pjson

Thanks for any pointers!

Regards,

Richard Duivenvoorde