Log message:
(geography/R-sf) 1.0.13 to 1.0.19

# version 1.0-19

* fix type checks in C++ GDAL area and length computation functions,
  anticipating GDAL 3.10.0; #2466, #2468, #2469 by @rsbivand and
  @rouault

* improve test on empty geometries, which changed in 1.0-18; #2463

* `gdal_utils()` `ogrinfo` has an argument `read_only` which, when
  `TRUE` (or `options` includes `"-ro"`), opens a datasource in
  read-only mode (#2460; `sf` did this before 1.0-17); by default a
  datasource is opened in update (read-write) mode (since sf 1.0-17;
  #2420)

* the `sf` -> `ppp` conversion `as.ppp.sf()` accepts a data.frame of
  marks instead of just 1 column #2450, by @agila5

* add flag for preservation of point order in `st_voronoi` #1371 for
  GEOS >= 3.12

# version 1.0-18

* support `POLYGON FULL` simple feature geometry, representing the
  entire Earth surface, as used by `s2geometry`; see also
  https://r-spatial.org/r/2024/10/11/polygonfull.html for a longer
  introduction; #2441

* `st_sfc()` has an argument `oriented` which, when set to `TRUE`,
  adds an attribute `oriented=TRUE` to an `sfc` object, indicating
  that this object should not be reoriented in conversion to
  `s2_geography` (avoiding using the global option `s2_oriented`);
  `st_as_sfc.bbox()` sets this to `TRUE`; #2441

* fix build failure with GDAL < 3.4.0 #2436

* `st_simplify()` now accepts feature-wise tolerance values when `s2`
  is switched on #2442

# version 1.0-17

* add `st_transform()` method for `bbox` objects; this uses
  OGRCoordinateTransformation::TransformBounds(), densifying first and
  antemeridian proof; #2415

* `st_filter.sf()` correctly scopes `x` and `y` arguments using !!
  operator; #2416

* `[.sfc` and `[<-.sfc` use matrix/array type subsetting for `sfc`
  objects that have a `dim` attribute

* add `st_exterior_ring()` to extract exterior rings (remove holes);
  #2406

* add `text.sf()`, `text.sfc()`, `points.sf()`, `points.sfc()` to
  annotate base plots at geometry centroids; #2399

* `st_sf()` no longer strips `tbl` or `tbl_df` class labels; #2378

* `st_layers()` returns an object of class `c("sf_layers",
  "data.frame")`, with a dedicated `print` method.

* when `dim` is not `XYZM`, `sf_as_sf.data.frame()` interprets a
  length 4 `coords` argument to specify the corners of a rectangular
  polygon; #2357

* `st_interpolate_aw()` gains an `na.rm` argument, for removing
  features with `NA` attributes before interpolating; #830

* `merge.sf()` no longer renames geometry column; #2334

# version 1.0-16

* `st_join()` no longer renames the geometry column; #2327

* `st_sample()` works when unprojected polygon geometry crosses the
  antemeridian; #2331

* clean up and modernization of S3 registration of methods and tests;
  #2285, #2288, #2316, #2341, #2342, by @olivroy

* `[.sfc` works when setting argument `op`; #2320

* `st_sample()` for polygons is sensitive to setting `oriented = TRUE`
  to prevent wrongly correcting ring directions; #2308

* add support for the GDAL `footprint` utility (requiring GDAL >=
  3.8.0) to `gdal_utils`; #2305, by @goergen95

* existing environment variables `PROJ_LIB` and `PROJ_DATA` are
  (again) ignored on `sf` binary CRAN installations (win + macos),
  effectively by overwriting them during the R session and restoring
  them on exit; this does not happen if environment variable
  `R_SF_USE_PROJ_DATA` is set to `true`. #2298

* add `st_line_project()` to find how far a point is when projected on
  a line; #2291

* add `st_line_interpolate()` to obtain a point at a certain distance
  along a line; #2291

# version 1.0-15

* add `st_perimeter()` generic to cover both geographic and projected
  coordinates; #268, #2279, by @JosiahParry

* add `st_sample()` method for `bbox`, with special provisions for
  ellipsoidal coordinates; #2283

* documentation clean-up by @olivroy; #2266, #2285

* `st_convex_hull()` uses `s2::s2_convex_hull()` for geodetic
  coordinates; #2250

* add `directed` argument to `st_line_merge()`; #2264

* `st_union.sfc()` given `x` and `y` works consistently across
  geodetic and projected objects; #2262

* `st_union.sf()` given `x` and `y` unions pairwise if `by_feature = TRUE`; #2259

* `st_read()` work around issue with GPKG driver if `wkt_filter` is set; #2248

* `st_read()` uses GDAL's stream reading when `use_stream = TRUE`;
  #2238 by @paleolimbot

* `st_transform()` responds to in-session changes to `sf_proj_network()`; #2166

* `plot.sf()`: `key.width` is sensitive to pointsize graphics
  parameter `par("ps")`; keys with factor levels suggest a proper size
  if they won't fit.

* `plot.sf()`: `key.pos` can hold a second value in [0, 1] determining
  the relative position of the key in the available space

* `[<-.sf` fixes the `agr` attribute when it is broken; #2211

* if the env. variable `ADD_SF_NAMESPACE` is set to `true`, `sf`
  objects get a new attribute, `.sf_namespace`, which forces loading
  the `sf` namespace when it has not been loaded so far, e.g. for
  proper printing or plotting of an `sf` object; #2212 by Mike Mahoney

* `distinct.sf()` is type-safe for `sf` objects with zero rows; #2204

* `summarise.sf()` raises an error if `.by` is given but no `across()`
  on the geometry; #2207

* `st_write()` matches fields on name first, than on position; this
  matters for formats that have pre-defined names, such as GPX; #2202

# version 1.0-14

* fix `plot.sf()` when using a key for multiple factor variables; #2196, #2195

* fix use of `as.numeric_version` in a test, for upcoming change in r-devel

* code tidy-ing: fix many lintr suggestions, thanks to Michael Chirico
  (#2181 - #2191)

Log message:
*: recursive bump for icu 76 shlib major version bump

Log message:
*: revbump for gdal

Log message:
*: revbump for icu downgrade

Log message:
*: recursive bump for icu 76.1 shlib bump

Log message:
*: revbump for proj update

less gdal, postgis, due to in-flight updates, to be handled separately

Log message:
revbump after icu and protobuf updates

Log message:
*: revbump for gdal-lib