leijurv left a comment (osm2pgsql-dev/osm2pgsql#2482)
> The mix of tile expiry and recursive query is interesting. I actually think
> that the solution is simpler because of your use of tile-based expiry instead
> of way-based. I did an implementation with way-based expiry for
> [waymarkedtrails](https://github.com/waymarkedtrails/osgende/blob/master/osgende/lines/segments.py)
> and only got it to work by tracking the IDs of the source ways for each
> merged way. That would make this implementation significantly more complex.
I admit I do not know how osm2pgsql works and clearly you know more than me
But this is my plan:
Right now we find deleted ways, updated ways, and created ways, by "rounding
up" to a Z/X/Y tile. i.e. we get a "low resolution picture" of where edits have
happened.
Instead, I'm imagining saving in the exact same way, at the exact same time,
expiry data that is simply more precise: the exact `ST_X(ST_StartPoint(way)),
ST_Y(ST_StartPoint(way))` and `ST_X(ST_EndPoint(way)), ST_Y(ST_EndPoint(way))`
(including for deleted ways!). In my mind, the queries in this PR would be
exactly(?) the same, only higher precision because they are a collection of
exact points, rather than polygon geometries. Like `_glm_region` is a
collection of `ST_Point` rather than a collection of tile bboxes. Possibly I am
missing something?
--
Reply to this email directly or view it on GitHub:
https://github.com/osm2pgsql-dev/osm2pgsql/pull/2482#issuecomment-4634591346
You are receiving this because you are subscribed to this thread.
Message ID: <osm2pgsql-dev/osm2pgsql/pull/2482/[email protected]>
_______________________________________________
Tile-serving mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/tile-serving
