Edzer Pebesma comments

Results 708 comments of


                                            Edzer Pebesma

st_sf fails to create SF with multiple SFC columns if are not empty

This change led to so many problems elsewhere that I have to revert it.

st_wrap_dateline() not working as expected

With ```r > library(sf) Linking to GEOS 3.12.1, GDAL 3.8.3, PROJ 9.3.1; sf_use_s2() is TRUE ``` I get ![x](https://github.com/r-spatial/sf/assets/520851/4cbdd9a0-6ffa-4dd8-afd7-0c830673b205) which seems the right thing.

st_make_valid() shifts coordinates

Please attach the file you used.

st_make_valid() shifts coordinates

Yes, after setting `sf_use_s2(FALSE)` you also get identical results.

st_make_valid() shifts coordinates

Digging into this I got the feeling that something happens inside `wk`: ```r library(sf) # Linking to GEOS 3.11.1, GDAL 3.6.4, PROJ 9.1.1; sf_use_s2() is TRUE vec

Autoconf script will break if proj version jumps to double digit

Great thinking - thanks for the PR!

Autoconf script will break if proj version jumps to double digit

Still fails for [wasm](https://github.com/r-universe/r-spatial/actions/runs/7477676230/job/20351060311) - any ideas what to do?

Autoconf script will break if proj version jumps to double digit

That now passes, but hangs on geos; I looked at the successful wasm build of sf, but could not find the configure step being executed there (which has the same...

mutate(.keep = "none") and mutate(.keep = "used") error with sf object

Thanks, FYI, > bind_cols() seems to work around the error, but looses the geometry; this may hint at what's happening. `bind_cols()` is a function, not a method, so there's little...

improved st_intersection(polygons) suggestion

Thanks; I guess you need this because `st_intersection()` doesn't do it for your case. Could you share a dataset that fails with `st_intersection` but succeeds with your code?