joto left a comment (osm2pgsql-dev/osm2pgsql#2427)
We should split the multipolygons first for the creation of expire tiles. But
the current code is somewhat complex and doesn't allow that easily and we need
to deal with the backwards compatibility issue that currently we are using the
whole bounding box to determine the cutoff in the hybrid case. We also should
avoid calculating the bounding box twice (once on the whole mp, once for all
the polygons). All of this is easier to reason about if we clean up the code
first.
Splitting into the mps would fix most of the cases where the current code
doesn't work, but not all of them. We do have at least one mp that goes 360°
[here](https://www.openstreetmap.org/relation/2186646) that this code will
break. I couldn't find a single case where this code acts as it was intended to
do, so I believe we should just remove it.
There is the separate issue of how to handle too large expire areas as
discussed [here](https://github.com/osm2pgsql-dev/osm2pgsql/discussions/2428)
but that's a separate issue. The current code saves us from expiring the whole
planet but still allows half the planet which is still way too big to be
sensible.
--
Reply to this email directly or view it on GitHub:
https://github.com/osm2pgsql-dev/osm2pgsql/pull/2427#issuecomment-3473166357
You are receiving this because you are subscribed to this thread.
Message ID: <osm2pgsql-dev/osm2pgsql/pull/2427/[email protected]>
_______________________________________________
Tile-serving mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/tile-serving
