Note to self: Between #QGIS versions 3.34.5 and 3.34.6, the packaging for the #Windows version of QGIS changed from #Python 3.9 to Python 3.12.

Will have to test if/how #Trajectools is affected

HT: https://gitlab.com/Oslandia/qgis/qduckdb by @oslandia

Oslandia / QGIS / QDuckDB Β· GitLab

Documentation: https://oslandia.gitlab.io/qgis/qduckdb/

GitLab
@underdarkGIS @oslandia Yes, and that should definitely not have happened ! Changing a major version of critical Python dependency in a minor QGIS release, without any warning, is totally nuts. There should have been a QEP for that, and at least clear communication to plugin developers.
@vpicavet yes, bit strange, but I might have missed messages. There's still this open issue, for example, https://github.com/qgis/QGIS/issues/54491
Upgrade default QGIS python from 3.9 to latest stable version Β· Issue #54491 Β· qgis/QGIS

Feature description Dear QGIS Team, I hope this message finds you well. I am writing to request an upgrade of the default Python version used in QGIS from 3.9 to the latest stable version, which is...

GitHub
@vpicavet @underdarkGIS @oslandia The issue is that OSGeo4W doesn't use dependency pinning or channels. So there's no way to update for let's say the current QGIS dev version without affecting the LTR as well. With Conda, you could have a dedicated qgis-ltr channel with its own set of dependencies.

@EvenRouault @vpicavet @underdarkGIS @oslandia

You could have a python3-core and python3-core-dev to deal with that issue, like gdal which has gdal 3.9 (targeted by qgis ltr) and gdal 3.11 (gdal-dev) targeted by qgis-dev.

Still, I agree OSGeo4W is IMHO to limited to tackle this kind of issues. But I would prefer vcpkg over conda