You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I guess this could be fixed by adding the relevant dispatches to ArchGDAL in which we convert the tuple to a point which might be fine for this application? (The more general question of "how to handle these functions when combining different GeoInterface enabled packages" feels fairly complicated...)
The text was updated successfully, but these errors were encountered:
To your second point, we're writing https://www.github.com/asinghvi17/GeometryOps.jl which is meant to provide Julia-native algorithms and a unified API for these kinds of geometry operations.
In general the dispatch could be added, but then you would get into adding a dispatch for each combination which seems a bit difficult...
Honestly I have pretty much given up on GeoInterface.distance and similar methods being something that can work in practice. We should really deprecate them, but in this case I didn't intend for returning a point to break them (ArchGDAL objects is one of the only use cases where they actually worked).
One easy fix is to return another Point type that just wraps the tuple but ArchGDAL owns and can dispatch on. Anything is better than the old C object Point that needs a finaliser (imagine having a million of those...).
But I pretty much stalled on fixing ArchGDAL geoms because its so hard to work on and after everything it is still really slow, especially without the point fix that caused this problem.
But, we should return a struct point instead of a Tuple. If you want to PR its probably only 10 lines.
I have some code which broke when the return type of
getgeom
changed in #369.A mwe of the problem would be:
I guess this could be fixed by adding the relevant dispatches to
ArchGDAL
in which we convert the tuple to a point which might be fine for this application? (The more general question of "how to handle these functions when combining differentGeoInterface
enabled packages" feels fairly complicated...)The text was updated successfully, but these errors were encountered: