Edzer Pebesma

Thanks; we also see now that the [windows GA](https://github.com/r-spatial/sf/runs/6169412990?check_suite_focus=true) breaks on having R 4.2 but not the required / corresponding winlibs things.

Dear package author, while checking reverse dependency's against the sf 1.0 candidate which uses s2 as a geometry back-end when using ellipsoidal coordinates, we found a regression. Please verify that...

For those interested in a low-res world map valid on S2, created by @paleolimbot : ```r library(sf) # Linking to GEOS 3.9.0, GDAL 3.2.1, PROJ 7.2.1 library(s2) st_as_sf(s2_data_tbl_countries) # Simple...

We can't call `s2_rebuild` when the error already happens in `s2_geog_from_wkb`: ```r library(tmap) library(sf) # Linking to GEOS 3.9.0, GDAL 3.2.1, PROJ 7.2.1 library(s2) data(World) x = st_transform(World, 4326) #...

And the following also didn't help: ```r w = s2_geog_from_wkb(st_as_binary(st_segmentize(st_geometry(x), units::set_units(10, km)))) w = s2_geog_from_wkb(st_as_binary(st_segmentize(st_geometry(x), units::set_units(1, km)))) w = s2_geog_from_wkb(st_as_binary(st_segmentize(st_geometry(x), units::set_units(.1, km)))) g = st_geometry(x) st_crs(g) = NA # so...

Bingo! This makes `tmap` and `tmaptools` pass check, and possibly a number of others. I will run an entire revdep check with this change, and report back. Out of interest:...

> I think it might be worth exposing check as an argument to st_as_s2() so that when a user gets this error they have recourse (rather than turning this on...

- [x] Use the `st_precision()` of a sfc to decide (i.e., if a precision is defined, export to WKB without rounding, read into s2 with `check = FALSE`, and call...

Hi @luukvdmeer ! I don't think you have to change anything of this kind, `sf` should do _all_ the conversion to and from `s2`, no need to change your package.

Fun fact: https://github.com/r-spatial/sf/commit/6de34206a19b3f33b939955bc4375de4756d5e83 would solve _all_ issues above. Too bad we can't go with it...