[R-package] Add support for R 4.0 (fixes #3064, fixes #3024)

.appveyor.yml

@@ -15,8 +15,12 @@ environment:
   matrix:
     - COMPILER: MINGW
       TASK: r-package
-    - COMPILER: MSVC
+      R_VERSION: 3.6
+      TOOLCHAIN: MINGW
+    - COMPILER: MSVC 
       TASK: r-package
+      R_VERSION: 4.0
+      TOOLCHAIN: MSVC
       APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2019
     - COMPILER: MSVC
       TASK: python

.ci/test_r_package.sh

@@ -7,6 +7,26 @@ mkdir -p $R_LIB_PATH
 echo "R_LIBS=$R_LIB_PATH" > ${HOME}/.Renviron
 export PATH="$R_LIB_PATH/R/bin:$PATH"
 
+# Get details needed for installing R components
+#
+# NOTES:
+#    * Linux builds on Azure use a container and don't need these details
+if ! { [[ $AZURE == "true" ]] && [[ $OS_NAME == "linux" ]]; }; then
+    R_MAJOR_VERSION=( ${R_VERSION//./ } )
+    if [[ "${R_MAJOR_VERSION}" == "3" ]]; then
+        export R_MAC_VERSION=3.6.3
+        export R_LINUX_VERSION="3.6.3-1bionic"
+        export R_APT_REPO="bionic-cran35/"
+    elif [[ "${R_MAJOR_VERSION}" == "4" ]]; then
+        export R_MAC_VERSION=4.0.0
+        export R_LINUX_VERSION="4.0.0-1.1804.0"
+        export R_APT_REPO="bionic-cran40/"
+    else
+        echo "Unrecognized R version: ${R_VERSION}"
+        exit -1
+    fi
+fi
+
 # installing precompiled R for Ubuntu
 # https://cran.r-project.org/bin/linux/ubuntu/#installation
 # adding steps from https://stackoverflow.com/a/56378217/3986677 to get latest version
@@ -18,7 +38,7 @@ if [[ $AZURE != "true" ]] && [[ $OS_NAME == "linux" ]]; then
         --keyserver keyserver.ubuntu.com \
         --recv-keys E298A3A825C0D65DFD57CBB651716619E084DAB9
     sudo add-apt-repository \
-        "deb https://cloud.r-project.org/bin/linux/ubuntu bionic-cran35/"
+        "deb https://cloud.r-project.org/bin/linux/ubuntu ${R_APT_REPO}"
     sudo apt-get update
     sudo apt-get install \
         --no-install-recommends \

.ci/test_r_package_windows.ps1

@@ -29,34 +29,60 @@ function Download-Miktex-Setup {
     Download-File-With-Retries $FileToDownload $destfile
 }
 
-$env:R_WINDOWS_VERSION = "3.6.3"
 $env:R_LIB_PATH = "$env:BUILD_SOURCESDIRECTORY/RLibrary" -replace '[\\]', '/'
 $env:R_LIBS = "$env:R_LIB_PATH"
-$env:PATH = "$env:R_LIB_PATH/Rtools/bin;" + "$env:R_LIB_PATH/R/bin/x64;" + "$env:R_LIB_PATH/miktex/texmfs/install/miktex/bin/x64;" + $env:PATH
+$env:PATH = "$env:R_LIB_PATH/Rtools/bin;" + "$env:R_LIB_PATH/Rtools/usr/bin;" + "$env:R_LIB_PATH/R/bin/x64;" + "$env:R_LIB_PATH/miktex/texmfs/install/miktex/bin/x64;" + $env:PATH
 $env:CRAN_MIRROR = "https://cloud.r-project.org/"
 $env:CTAN_MIRROR = "https://ctan.math.illinois.edu/systems/win32/miktex"
 $env:CTAN_MIKTEX_ARCHIVE = "$env:CTAN_MIRROR/setup/windows-x64/"
 $env:CTAN_PACKAGE_ARCHIVE = "$env:CTAN_MIRROR/tm/packages/"
 
+# Get details needed for installing R components
+#
+# NOTES:
+#    * some paths and file names are different on R4.0
+$env:R_MAJOR_VERSION = $env:R_VERSION.split('.')[0]
+if ($env:R_MAJOR_VERSION -eq "3") {
+  $env:RTOOLS_MINGW_BIN = "$env:R_LIB_PATH/Rtools/mingw_64/bin"
+  $env:RTOOLS_EXE_FILE = "Rtools35.exe"
+  $env:R_WINDOWS_VERSION = "3.6.3"
+} elseif ($env:R_MAJOR_VERSION -eq "4") {
+  $env:RTOOLS_MINGW_BIN = "$env:R_LIB_PATH/Rtools/mingw64/bin"
+  $env:RTOOLS_EXE_FILE = "rtools40-x86_64.exe"
+  $env:R_WINDOWS_VERSION = "4.0.0"
+} else {
+  Write-Output "[ERROR] Unrecognized R version: $env:R_VERSION"
+  Check-Output $false
+}
+
 if ($env:COMPILER -eq "MINGW") {
-  $env:CXX = "$env:R_LIB_PATH/Rtools/mingw_64/bin/g++.exe"
-  $env:CC = "$env:R_LIB_PATH/Rtools/mingw_64/bin/gcc.exe"
+  $env:CXX = "$env:RTOOLS_MINGW_BIN/g++.exe"
+  $env:CC = "$env:RTOOLS_MINGW_BIN/gcc.exe"
 }
 
 cd $env:BUILD_SOURCESDIRECTORY
 tzutil /s "GMT Standard Time"
 [Void][System.IO.Directory]::CreateDirectory($env:R_LIB_PATH)
 
-if ($env:COMPILER -eq "MINGW") {
+if ($env:TOOLCHAIN -eq "MINGW") {
   Write-Output "Telling R to use MinGW"
   $install_libs = "$env:BUILD_SOURCESDIRECTORY/R-package/src/install.libs.R"
-  ((Get-Content -path $install_libs -Raw) -replace 'use_mingw <- FALSE','use_mingw <- TRUE') | Set-Content -Path $install_libs
+  ((Get-Content -Path $install_libs -Raw) -Replace 'use_mingw <- FALSE','use_mingw <- TRUE') | Set-Content -Path $install_libs
+} elseif ($env:TOOLCHAIN -eq "MSYS") {
+  Write-Output "Telling R to use MSYS"
+  $install_libs = "$env:BUILD_SOURCESDIRECTORY/R-package/src/install.libs.R"
+  ((Get-Content -Path $install_libs -Raw) -Replace 'use_msys2 <- FALSE','use_msys2 <- TRUE') | Set-Content -Path $install_libs
+} elseif ($env:TOOLCHAIN -eq "MSVC") {
+  # no customization for MSVC
+} else {
+  Write-Output "[ERROR] Unrecognized compiler: $env:TOOLCHAIN"
+  Check-Output $false
 }
 
 # download R and RTools
 Write-Output "Downloading R and Rtools"
 Download-File-With-Retries -url "https://cloud.r-project.org/bin/windows/base/old/$env:R_WINDOWS_VERSION/R-$env:R_WINDOWS_VERSION-win.exe" -destfile "R-win.exe"
-Download-File-With-Retries -url "https://cloud.r-project.org/bin/windows/Rtools/Rtools35.exe" -destfile "Rtools.exe"
+Download-File-With-Retries -url "https://cloud.r-project.org/bin/windows/Rtools/$env:RTOOLS_EXE_FILE" -destfile "Rtools.exe"
 
 # Install R
 Write-Output "Installing R"
@@ -67,9 +93,9 @@ Write-Output "Installing Rtools"
 Start-Process -FilePath Rtools.exe -NoNewWindow -Wait -ArgumentList "/VERYSILENT /DIR=$env:R_LIB_PATH/Rtools" ; Check-Output $?
 Write-Output "Done installing Rtools"
 
-# MiKTeX and pandoc can be skipped on non-MINGW builds, since we don't
+# MiKTeX and pandoc can be skipped on MSVC builds, since we don't
 # build the package documentation for those
-if ($env:COMPILER -eq "MINGW") {
+if ($env:COMPILER -ne "MSVC") {
     Download-Miktex-Setup "$env:CTAN_MIKTEX_ARCHIVE" "miktexsetup-x64.zip"
     Add-Type -AssemblyName System.IO.Compression.FileSystem
     [System.IO.Compression.ZipFile]::ExtractToDirectory("miktexsetup-x64.zip", "miktex")
@@ -115,7 +141,7 @@ if ($env:COMPILER -ne "MSVC") {
   $note_str = Get-Content "${LOG_FILE_NAME}" | Select-String -Pattern ' NOTE' | Out-String ; Check-Output $?
   $relevant_line = $note_str -match '.*Status: (\d+) NOTE.*'
   $NUM_CHECK_NOTES = $matches[1]
-  $ALLOWED_CHECK_NOTES = 3
+  $ALLOWED_CHECK_NOTES = 4
   if ([int]$NUM_CHECK_NOTES -gt $ALLOWED_CHECK_NOTES) {
       Write-Output "Found ${NUM_CHECK_NOTES} NOTEs from R CMD check. Only ${ALLOWED_CHECK_NOTES} are allowed"
       Check-Output $False
@@ -139,6 +165,16 @@ if ($checks.Matches.length -eq 0) {
   Check-Output $False
 }
 
+# Checking that we got the right toolchain for MinGW. If using MinGW, both
+# MinGW and MSYS toolchains are supported
+if ($env:COMPILER -eq "MINGW") {
+  $checks = Select-String -Path "${INSTALL_LOG_FILE_NAME}" -Pattern "Trying to build with.*$env:TOOLCHAIN"
+  if ($checks.Matches.length -eq 0) {
+    Write-Output "The wrong toolchain was used. Check the build logs."
+    Check-Output $False
+  }
+}
+
 if ($env:COMPILER -eq "MSVC") {
   Write-Output "Running tests with testthat.R"
   cd R-package/tests

.github/workflows/main.yml

@@ -10,7 +10,7 @@ on:
 
 jobs:
   test:
-    name: ${{ matrix.task }} (${{ matrix.os }}, ${{ matrix.compiler }})
+    name: ${{ matrix.task }} (${{ matrix.os }}, ${{ matrix.compiler }}, R ${{ matrix.r_version }})
     runs-on: ${{ matrix.os }}
     strategy:
       fail-fast: false
@@ -19,15 +19,35 @@ jobs:
           - os: ubuntu-latest
             task: r-package
             compiler: gcc
+            r_version: 3.6
+          - os: ubuntu-latest
+            task: r-package
+            compiler: gcc
+            r_version: 4.0
+          - os: ubuntu-latest
+            task: r-package
+            compiler: clang
+            r_version: 3.6
           - os: ubuntu-latest
             task: r-package
             compiler: clang
+            r_version: 4.0
           - os: macOS-latest
             task: r-package
             compiler: gcc
+            r_version: 3.6
+          - os: macOS-latest
+            task: r-package
+            compiler: gcc
+            r_version: 4.0
+          - os: macOS-latest
+            task: r-package
+            compiler: clang
+            r_version: 3.6
           - os: macOS-latest
             task: r-package
             compiler: clang
+            r_version: 4.0
     steps:
       - name: Checkout repository
         uses: actions/checkout@v1
@@ -53,5 +73,6 @@ jobs:
           export CONDA="$HOME/miniconda"
           export PATH="$CONDA/bin:${HOME}/.local/bin:$PATH"
           export LGB_VER=$(head -n 1 VERSION.txt)
+          export R_VERSION="${{ matrix.r_version }}"
           $GITHUB_WORKSPACE/.ci/setup.sh
           $GITHUB_WORKSPACE/.ci/test.sh

.gitignore

@@ -406,6 +406,7 @@ R-package/src-i386
 lightgbm_r/*
 lightgbm*.tar.gz
 lightgbm.Rcheck/
+*.def
 
 # Files created by R examples and tests
 **/lgb-Dataset.data

.vsts-ci.yml

@@ -109,11 +109,18 @@ jobs:
   pool:
     vmImage: 'vs2017-win2016'
   strategy:
-    maxParallel: 4
+    maxParallel: 5
     matrix:
-      r_package:
+      r_package_msvc:
         TASK: r-package
         COMPILER: MSVC
+        R_VERSION: 3.6
+        TOOLCHAIN: MSVC
+      r_package_msys:
+        TASK: r-package
+        COMPILER: MINGW
+        R_VERSION: 4.0
+        TOOLCHAIN: MSYS
       regular:
         TASK: regular
         PYTHON_VERSION: 3.6

R-package/README.md

@@ -20,14 +20,53 @@ Note: 32-bit (i386) R/Rtools is currently not supported.
 
 #### Windows Preparation
 
-Installing [Rtools](https://cran.r-project.org/bin/windows/Rtools/) is mandatory, and only support the 64-bit version. It requires to add to PATH the Rtools MinGW64 folder, if it was not done automatically during installation.
+Installing a 64-bit version of [Rtools](https://cran.r-project.org/bin/windows/Rtools/) is mandatory.
 
-The default compiler is Visual Studio (or [VS Build Tools](https://visualstudio.microsoft.com/downloads/)) in Windows, with an automatic fallback to Rtools or any [MinGW64](https://sourceforge.net/projects/mingw-w64/files/Toolchains%20targetting%20Win64/Personal%20Builds/mingw-builds/) (x86_64-posix-seh) available (this means if you have only Rtools and CMake, it will compile fine).
+After installing `Rtools` and `CMake`, be sure the following paths are added to the environment variable `PATH`. These may have been automatically added when installing other software.
 
-To force the usage of Rtools / MinGW, you can set `use_mingw` to `TRUE` in `R-package/src/install.libs.R`.
+* `Rtools`
+    - If you have `Rtools` 3.x, example:
+        - `C:\Rtools\mingw_64\bin`
+    - If you have `Rtools` 4.0, example:
+        - `C:\rtools40\mingw64\bin`
+        - `C:\rtools40\usr\bin`
+* `CMake`
+    - example: `C:\Program Files\CMake\bin`
+* `R`
+    - example: `C:\Program Files\R\R-3.6.1\bin`
+
+NOTE: Two `Rtools` paths are required from `Rtools` 4.0 onwards because paths and the list of included software was changed in `Rtools` 4.0.
+
+#### Windows Toolchain Options
+
+A "toolchain" refers to the collection of software used to build the library. The R package can be built with three different toolchains.
 
 **Warning for Windows users**: it is recommended to use *Visual Studio* for its better multi-threading efficiency in Windows for many core systems. For very simple systems (dual core computers or worse), MinGW64 is recommended for maximum performance. If you do not know what to choose, it is recommended to use [Visual Studio](https://visualstudio.microsoft.com/downloads/), the default compiler. **Do not try using MinGW in Windows on many core systems. It may result in 10x slower results than Visual Studio.**
 
+**Visual Studio (default)**
+
+By default, the package will be built with [Visual Studio Build Tools](https://visualstudio.microsoft.com/downloads/).
+
+**MinGW (R 3.x)**
+
+If you are using R 3.x and installation fails with Visual Studio, `LightGBM` will fall back to using [MinGW](http://mingw-w64.org/doku.php) bundled with `Rtools`.
+
+If you want to force `LightGBM` to use MinGW (for any R version), open `R-package/src/install.libs.R` and change `use_mingw`:
+
+```r
+use_mingw <- TRUE
+```
+
+**MSYS2 (R 4.x)**
+
+If you are using R 4.x and installation fails with Visual Studio, `LightGBM` will fall back to using [MSYS2](https://www.msys2.org/). This should work with the tools already bundled in `Rtools` 4.0.
+
+If you want to force `LightGBM` to use MSYS2 (for any R version), open `R-package/src/install.libs.R` and change `use_msys2`:
+
+```r
+use_msys2 <- TRUE
+```
+
 #### Mac OS Preparation
 
 You can perform installation either with **Apple Clang** or **gcc**. In case you prefer **Apple Clang**, you should install **OpenMP** (details for installation can be found in [Installation Guide](https://github.com/microsoft/LightGBM/blob/master/docs/Installation-Guide.rst#apple-clang)) first and **CMake** version 3.16 or higher is required. In case you prefer **gcc**, you need to install it (details for installation can be found in [Installation Guide](https://github.com/microsoft/LightGBM/blob/master/docs/Installation-Guide.rst#gcc)) and set some environment variables to tell R to use `gcc` and `g++`. If you install these from Homebrew, your versions of `g++` and `gcc` are most likely in `/usr/local/bin`, as shown below.

R-package/inst/make-r-def.R

@@ -0,0 +1,98 @@
+# [description]
+#     Create a definition file (.def) from a .dll file, using objdump.
+#
+# [usage]
+#
+#     Rscript make-r-def.R something.dll something.def
+#
+# [references]
+#    * https://www.cs.colorado.edu/~main/cs1300/doc/mingwfaq.html
+
+args <- commandArgs(trailingOnly = TRUE)
+
+IN_DLL_FILE <- args[[1L]]
+OUT_DEF_FILE <- args[[2L]]
+DLL_BASE_NAME <- basename(IN_DLL_FILE)
+
+message(sprintf("Creating '%s' from '%s'", OUT_DEF_FILE, IN_DLL_FILE))
+
+# system() will not raise an R exception if the process called
+# fails. Wrapping it here to get that behavior.
+#
+# system() introduces a lot of overhead, at least on Windows,
+# so trying processx if it is available
+.pipe_shell_command_to_stdout <- function(command, args, out_file) {
+    has_processx <- suppressMessages({
+      suppressWarnings({
+        require("processx")  # nolint
+      })
+    })
+    if (has_processx) {
+        p <- processx::process$new(
+            command = command
+            , args = args
+            , stdout = out_file
+            , windows_verbatim_args = FALSE
+        )
+        invisible(p$wait())
+    } else {
+        message(paste0(
+          "Using system2() to run shell commands. Installing "
+          , "'processx' with install.packages('processx') might "
+          , "make this faster."
+        ))
+        # shQuote() is necessary here since one of the arguments
+        # is a file-path to R.dll, which may have spaces. processx
+        # does such quoting but system2() does not
+        exit_code <- system2(
+            command = command
+            , args = shoQuote(args)
+            , stdout = out_file
+        )
+        if (exit_code != 0L) {
+            stop(paste0("Command failed with exit code: ", exit_code))
+        }
+    }
+    return(invisible(NULL))
+}
+
+# use objdump to dump all the symbols
+OBJDUMP_FILE <- "objdump-out.txt"
+.pipe_shell_command_to_stdout(
+    command = "objdump"
+    , args = c("-p", IN_DLL_FILE)
+    , out_file = OBJDUMP_FILE
+)
+
+objdump_results <- readLines(OBJDUMP_FILE)
+invisible(file.remove(OBJDUMP_FILE))
+
+# Only one table in the objdump results matters for our purposes,
+# see https://www.cs.colorado.edu/~main/cs1300/doc/mingwfaq.html
+start_index <- which(
+    grepl(
+        pattern = "[Ordinal/Name Pointer] Table"
+        , x = objdump_results
+        , fixed = TRUE
+    )
+)
+empty_lines <- which(objdump_results == "")
+end_of_table <- empty_lines[empty_lines > start_index][1L]
+
+# Read the contents of the table
+exported_symbols <- objdump_results[(start_index + 1L):end_of_table]
+exported_symbols <- gsub("\t", "", exported_symbols)
+exported_symbols <- gsub(".*\\] ", "", exported_symbols)
+exported_symbols <- gsub(" ", "", exported_symbols)
+
+# Write R.def file
+writeLines(
+    text = c(
+        paste0("LIBRARY \"", DLL_BASE_NAME, "\"")
+        , "EXPORTS"
+        , exported_symbols
+    )
+    , con = OUT_DEF_FILE
+    , sep = "\n"
+)
+message(sprintf("Successfully created '%s'", OUT_DEF_FILE))