Open edzer opened 1 month ago
That's a great point...the S2Builder doesn't know about full polygons (I seem to remember) and there are some special gymnastics to ensure that it is handled properly in boolean operations. A single full polygon can probably be special-cased...I am not sure what would happen if one tried to rebuild a polygon that contained a FULL loop, though.
I don't think we have an s2_is_full()
to special-case this at the R level but that would basically be the workaround.
s2_is_full()
would be great to have!
I'm (finally) implementing support for
POLYGON FULL
insf
, and ran intos2_rebuild
converting an s2 full polygon to aGEOMETRYCOLLECTION EMPTY
. I don't think it should! I'll work around it in sf, but it would be nice if it wasn't touched ins2_rebuild
: In a debug session insf::st_make_valid()
: