Skip to contents

gdalraster.windows

gdalraster.windows builds a self-contained gdalraster package against a modern GDAL runtime published by this repository.

The installed gdalraster vendors GDAL and dependency DLLs, matching GDAL/PROJ data, and osgeo_utils. Later sessions require no activation helper, .Rprofile hook, PATH edit, or GDAL environment variables.

Requirements

  • Windows with R.
  • Rtools matching R.

Setup

Install the package, then build a self-contained gdalraster into the regular user library:

pak::pak("jimbrig/gdalraster.windows")
gdalraster.windows::gdal_setup()

Then, in any fresh session:

Pass lib for a custom library, or isolated = TRUE for a package-managed library that does not replace .libPaths()[1].

Status, verification, and updates 

gdalraster.windows::gdal_sitrep()
gdalraster.windows::gdal_verify()
gdalraster.windows::gdal_update()

Verification runs in fresh processes and checks:

  • the GDAL Algorithm API registry;
  • Arrow, Parquet, HDF5, and netCDF drivers;
  • GEOS and CRS resolution;
  • a first Parquet dataset open; and
  • gdal driver gpkg validate when embedded Python is ready.

Offline installation 

gdalraster.windows::gdal_setup(
  local_zip = "C:/Downloads/gdal-bundle-v3.13.2-windows-x64.zip"
)

Bundle assets are published at https://github.com/jimbrig/gdalraster.windows/releases.

Design

The package uses the downloaded bundle as a build-time SDK. After compiling gdalraster, it stages and vendors:

gdalraster/
  libs/x64/                  # gdalraster.dll + GDAL dependency closure
  gdal/                      # matching GDAL data
  proj/                      # matching PROJ data
  python/osgeo_utils/        # embedded-Python algorithms
  gdalraster.windows-build.dcf

The SDK has its own MANIFEST.dcf. Comparing bundle tags lets gdal_sitrep() detect a stale build after a runtime update.

Embedded Python is provisioned through a managed .pth in system CPython’s site-packages; the package never sets PYTHONPATH.

Repository responsibilities

  • .github/workflows/build.yml, tools/build_gdal.sh, and tools/collect_dlls.sh build and publish the GDAL bundle.
  • The R package installs the SDK, builds and vendors gdalraster, provisions Python, and verifies the user workflow.
  • .github/workflows/e2e.yml proves fresh-session self-containment.
  • .github/workflows/edge-cases.yml exercises dirty-machine failure modes.

See the published Getting Started, Runtime Guide, Architecture, and Troubleshooting articles for details.