[ci] Fix CI job for R artifact

[StrikerRUS]

Fixed #4810.

DO NOT MERGE before uncommenting # condition: not(startsWith(variables['Build.SourceBranch'], 'refs/pull/')).

[StrikerRUS]

Compilation for Matrix fails but the whole job succeeds. I guess this is due to pre-installed Matrix (not directly but probably as a dependency)

...

make[2]: Leaving directory '/tmp/RtmpXPS3LD/R.INSTALLaec3e29fc26/Matrix/src/SuiteSparse_config'
gcc -std=gnu99 -std=gnu11 -shared -L/usr/lib/R/lib -Wl,-Bsymbolic-functions -Wl,-z,relro -o Matrix.so CHMfactor.o Csparse.o TMatrix_as.o Tsparse.o init.o Mutils.o chm_common.o cs.o cs_utils.o dense.o dgCMatrix.o dgTMatrix.o dgeMatrix.o dpoMatrix.o dppMatrix.o dsCMatrix.o dsyMatrix.o dspMatrix.o dtCMatrix.o dtTMatrix.o dtrMatrix.o dtpMatrix.o factorizations.o ldense.o lgCMatrix.o sparseQR.o abIndex.o CHOLMOD.a COLAMD.a AMD.a SuiteSparse_config.a -llapack -lblas -lgfortran -lm -lquadmath -L/usr/lib/R/lib -lR
/usr/bin/ld: cannot find -lgfortran
collect2: error: ld returned 1 exit status
make[1]: *** [/usr/share/R/share/make/shlib.mk:10: Matrix.so] Error 1
make[1]: Leaving directory '/tmp/RtmpXPS3LD/R.INSTALLaec3e29fc26/Matrix/src'
ERROR: compilation failed for package ‘Matrix’
* removing ‘/home/vsts_azpcontainer/Rlib/Matrix’

The downloaded source packages are in
	‘/tmp/RtmpDMyuEa/downloaded_packages’
Warning message:
In install.packages(c("R6", "data.table", "jsonlite", "knitr", "Matrix",  :
  installation of one or more packages failed,
  probably ‘Matrix’

...

Need use --r-executable=RD and RDscript?..

[StrikerRUS]

Seems that using dev versions of R and Rscript helped.

[StrikerRUS]

@jameslamb You can find resulted artifact here: https://dev.azure.com/lightgbm-ci/_apis/resources/Containers/6904809/R-package?itemPath=R-package%2Flightgbm-3.3.1.99-r-cran.tar.gz

[jameslamb]

Looks ok to me!

I'm comfortable using RD in the wch1/r-debug1 image because:

1. We're only creating a source package, so I'm not worried about an alternative R build generating different artifacts.
2. The package will be tested against r-devel when uploaded to CRAN anyway, so there's no risk of "created a package using a non-standard version of R that breaks things on a CRAN submission".

I pulled the artifact created from this PR's most recent CI run (#4811 (comment)) and ran R CMD check --as-cran on it, just on my Mac (R 4.1.1 from CRAN, macOS 11.14). No issues found.

So I think it's ok to merge this! (after reverting the commented-out condition line). Thanks for working through this, sorry I didn't think of it in #4775 .

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.