Skip to content

Latest commit

 

History

History

docker

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Build and check sf against r-release and r-devel

See here for general docker files building R with a significant number of spatial extension packages and external dependencies, based on rocker and debian.

To allows building sf in alternative environments with all external system requirements (udunits, proj, gdal, geos, lwgeom), this directory has subdirectories with Docker files:

  • base: for installing R-release with all system dependencies required by sf,
  • devel: for building R-devel from source (downloaded from svn, without X11) on top of that.
  • gdal: for testing with newer/newst gdal/geos/proj versions
  • custom: for testing with libraries (gdal, geos, proj.4) installed in custom, non-standard directories
  • lowest: for checking sf against the lowest required PROJ, GDAL and GEOS versions
  • fedora: for checking sf against the fedora/clang platform
  • cran: using the debian setup used by several of the CRAN linux servers

Except for the cran image, all images are built on ubuntu:16.04 (xenial). They use ubuntugis-unstable for GIS package system dependencies.

run check under R release

In directory base, type

docker build . -t sf

this will build a docker image of approx. 3.8 Gb size, called sf. It will also install package sf with its dependencies from CRAN, and run a check on the github sources.

Run a container from this image interactively with

docker run -ti sf

If you want to run it while mounting the current working directory (pwd) to /pkg inside the container, and want to remove the container after you exit, use:

docker run -v `pwd`:/pkg --rm -ti sf

build R-devel, check with R-devel

After you built docker image sf above, on top of that you can install r-devel; it counts up to 4 Gb. Build it by executing the following command in the devel directory:

docker build . -t sf_devel

Run a container from this image interactively with

docker run -ti sf_devel

to start R-devel in the container, use the Rdevel command. Building the image checks sf from github.

If you want to run R with gdb in a container, use e.g.

docker run --cap-add=SYS_PTRACE --security-opt seccomp=unconfined -ti gdal3

with custom gdal, geos and proj installs

  翻译: