[R-package] Use command-line args to customize CMake builds (fixes #2441) (#3574)

* [R-package] Use command-line args to customize CMake builds (fixes #2441)

* Apply suggestions from code review
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>
Co-authored-by: Nikita Titov <nekit94-08@mail.ru>

.ci/test_r_package_windows.ps1

@@ -71,22 +71,6 @@ cd $env:BUILD_SOURCESDIRECTORY
 tzutil /s "GMT Standard Time"
 [Void][System.IO.Directory]::CreateDirectory($env:R_LIB_PATH)
 
-if ($env:R_BUILD_TYPE -eq "cmake") {
-  $install_libs = "$env:BUILD_SOURCESDIRECTORY/R-package/src/install.libs.R"
-  if ($env:TOOLCHAIN -eq "MINGW") {
-    Write-Output "Telling R to use MinGW"
-    ((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"
-    ((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 toolchain: $env:TOOLCHAIN"
-    Check-Output $false
-  }
-}
-
 # download R and RTools
 Write-Output "Downloading R and Rtools"
 Download-File-With-Retries -url "https://cran.r-project.org/bin/windows/base/old/$env:R_WINDOWS_VERSION/R-$env:R_WINDOWS_VERSION-win.exe" -destfile "R-win.exe"
@@ -131,7 +115,19 @@ if ($env:COMPILER -ne "MSVC") {
   $LOG_FILE_NAME = "lightgbm.Rcheck/00check.log"
 
   if ($env:R_BUILD_TYPE -eq "cmake") {
-    Run-R-Code-Redirect-Stderr "commandArgs <- function(...){'--skip-install'}; source('build_r.R')"; Check-Output $?
+    if ($env:TOOLCHAIN -eq "MINGW") {
+      Write-Output "Telling R to use MinGW"
+      $env:BUILD_R_FLAGS = "c('--skip-install', '--use-mingw')"
+    } elseif ($env:TOOLCHAIN -eq "MSYS") {
+      Write-Output "Telling R to use MSYS"
+      $env:BUILD_R_FLAGS = "c('--skip-install', '--use-msys2')"
+    } elseif ($env:TOOLCHAIN -eq "MSVC") {
+      $env:BUILD_R_FLAGS = "'--skip-install'"
+    } else {
+      Write-Output "[ERROR] Unrecognized toolchain: $env:TOOLCHAIN"
+      Check-Output $false
+    }
+    Run-R-Code-Redirect-Stderr "commandArgs <- function(...){$env:BUILD_R_FLAGS}; source('build_r.R')"; Check-Output $?
   } elseif ($env:R_BUILD_TYPE -eq "cran") {
     Run-R-Code-Redirect-Stderr "result <- processx::run(command = 'sh', args = 'build-cran-package.sh', echo = TRUE, windows_verbatim_args = FALSE, error_on_status = TRUE)" ; Check-Output $?
     # Test CRAN source .tar.gz in a directory that is not this repo or below it.

R-package/README.md

@@ -115,20 +115,20 @@ By default, the package will be built with [Visual Studio Build Tools](https://v
 
 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`:
+If you want to force `LightGBM` to use MinGW (for any R version), pass `--use-mingw` to the installation script.
 
-```r
-use_mingw <- TRUE
+```shell
+Rscript build_r.R --use-mingw
 ```
 
 **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`:
+If you want to force `LightGBM` to use MSYS2 (for any R version), pass `--use-msys2` to the installation script.
 
-```r
-use_msys2 <- TRUE
+```shell
+Rscript build_r.R --use-msys2
 ```
 
 #### Mac OS Preparation
@@ -150,15 +150,24 @@ cd LightGBM
 Rscript build_r.R
 ```
 
-The `build_r.R` script builds the package in a temporary directory called `lightgbm_r`. It will destroy and recreate that directory each time you run the script.
+The `build_r.R` script builds the package in a temporary directory called `lightgbm_r`. It will destroy and recreate that directory each time you run the script. That script supports the following command-line options:
 
-Note: for the build with Visual Studio/VS Build Tools in Windows, you should use the Windows CMD or Powershell.
+- `--skip-install`: Build the package tarball, but do not install it.
+- `--use-gpu`: Build a GPU-enabled versioon of the library
+- `--use-mingw`: Force the use of MinGW toolchain, regardless of R version.
+- `--use-msys2`: Force the use of MSYS2 toolchain, regardless of R version.
+
+Note: for the build with Visual Studio/VS Build Tools in Windows, you should use the Windows CMD or PowerShell.
 
 ### Installing a GPU-enabled Build
 
-Set `use_gpu` to `TRUE` in `R-package/src/install.libs.R` to enable the build with GPU support. You will need to install Boost and OpenCL first: details for installation can be found in [Installation-Guide](https://github.com/microsoft/LightGBM/blob/master/docs/Installation-Guide.rst#build-gpu-version).
+You will need to install Boost and OpenCL first: details for installation can be found in [Installation-Guide](https://github.com/microsoft/LightGBM/blob/master/docs/Installation-Guide.rst#build-gpu-version).
+
+After installing these other libraries, follow the steps in ["Installing from Source with CMake"](#install). When you reach the step that mentions `build_r.R`, pass the flag `--use-gpu`.
 
-After installing these other libraries, follow the steps in ["Installing from Source with CMake"](#install).
+```shell
+Rscript build_r.R --use-gpu
+```
 
 ### Installing Precompiled Binaries
 

build_r.R

@@ -10,10 +10,42 @@ INSTALL_AFTER_BUILD <- !("--skip-install" %in% args)
 TEMP_R_DIR <- file.path(getwd(), "lightgbm_r")
 TEMP_SOURCE_DIR <- file.path(TEMP_R_DIR, "src")
 
+USING_GPU <- "--use-gpu" %in% args
+USING_MINGW <- "--use-mingw" %in% args
+USING_MSYS2 <- "--use-msys2" %in% args
+
+recognized_args <- c(
+  "--skip-install"
+  , "--use-gpu"
+  , "--use-mingw"
+  , "--use-msys2"
+)
+unrecognized_args <- setdiff(args, recognized_args)
+if (length(unrecognized_args) > 0L) {
+  msg <- paste0(
+    "Unrecognized arguments: "
+    , paste0(unrecognized_args, collapse = ", ")
+  )
+  stop(msg)
+}
+
+# [description] Replace statements in install.libs.R code based on
+#               command-line flags
+.replace_flag <- function(variable_name, value, content) {
+  out <- gsub(
+    pattern = paste0(variable_name, " <-.*")
+    , replacement = paste0(variable_name, " <- ", as.character(value))
+    , x = content
+  )
+  return(out)
+}
+
 install_libs_content <- readLines(
   file.path("R-package", "src", "install.libs.R")
 )
-USING_GPU <- any(grepl("use_gpu.*TRUE", install_libs_content))
+install_libs_content <- .replace_flag("use_gpu", USING_GPU, install_libs_content)
+install_libs_content <- .replace_flag("use_mingw", USING_MINGW, install_libs_content)
+install_libs_content <- .replace_flag("use_msys2", USING_MSYS2, install_libs_content)
 
 # R returns FALSE (not a non-zero exit code) if a file copy operation
 # breaks. Let's fix that
@@ -75,6 +107,12 @@ result <- file.copy(
 )
 .handle_result(result)
 
+# overwrite src/install.libs.R with new content based on command-line flags
+writeLines(
+  text = install_libs_content
+  , con = file.path(TEMP_SOURCE_DIR, "install.libs.R")
+)
+
 # Add blank Makevars files
 result <- file.copy(
   from = file.path(TEMP_R_DIR, "inst", "Makevars")