[R-package] suppress Wcast-function-type warning in CMake-based gcc and MinGW builds (fixes #4273)

[jameslamb]

This PR fixes #4273. It proposes suppressing this warning on CMake-based builds of the R package using gcc or MinGW:

warning: cast between incompatible function types from 'SEXPREC* (*)(LGBM_SE, SEXP, SEXP)' {aka 'SEXPREC* (*)(LGBM_SER*, SEXPREC*, SEXPREC*)'} to 'DL_FUNC' {aka 'void* (*)()'} [-Wcast-function-type]

This PR proposes suppressing this warning specifically for the portion of lightgbm_R.cpp that handles routine registration. The pattern used in {lightgbm} is exactly the same one used in many other R packages on CRAN (https://github.com/search?q=R_CallMethodDef+org%3Acran+filename%3A*.cpp), and the extensive unit tests covering the package give me confidence that this isn't causing any issues.

I think that removing this warning is useful because it remove unnecessary noise from users' logs and might prevent users who are trying to debug installations issues from spending time focused on warnings that are not actually problematic.

Notes for reviewers

See #4273 for more details on this warning.

See https://cran.r-project.org/doc/manuals/r-release/R-exts.html#useDynLib and the description in #4155 for more background information on how routine registration in R packages works.

CRAN does have a rule preventing the of pragmas that suppress warnings, but the pragmas introduced in this PR will be removed when preparing a CRAN package.

LightGBM/build-cran-package.sh

Line 94 in 08d1ce4

-e 's/^.*#pragma GCC diagnostic.*$//' \

[jameslamb]

I've converted this back to a draft. It seems R CMD check is unhappy with the use of this pragma.

All CMake-based builds from https://github.com/microsoft/LightGBM/pull/4274/checks?check_run_id=2550215720 are failing with this warning.

* checking pragmas in C/C++ headers and code ... WARNING

File which contains non-portable pragma(s)

  'src/src/lightgbm_R.cpp'

File which contains pragma(s) suppressing diagnostics:

  'src/src/lightgbm_R.cpp'

The MinGW + R 3.6 build (but not the MinGW + R 4.0 build) also has this warning:

D:\a\LightGBM\LightGBM\lightgbm.Rcheck\00_pkg_src\lightgbm\src\src\lightgbm_R.cpp:684:32: warning: unknown option after '#pragma GCC diagnostic' kind [-Wpragmas]

 #pragma GCC diagnostic ignored "-Wcast-function-type"

link: https://github.com/microsoft/LightGBM/pull/4274/checks?check_run_id=2550215720

and clang builds have this

/home/runner/work/LightGBM/LightGBM/lightgbm.Rcheck/00_pkg_src/lightgbm/src/src/lightgbm_R.cpp:684:32: warning: unknown warning group '-Wcast-function-type', ignored [-Wunknown-warning-option]
#pragma GCC diagnostic ignored "-Wcast-function-type"
                               ^
1 warning generated.

link: https://github.com/microsoft/LightGBM/pull/4274/checks?check_run_id=2550215599

So now I think that maybe it would be better to suppress this from the CMake side (but only for the R package).

[jameslamb]

Ok I think this has successfully suppressed this warning, and this PR is ready for review.

• gcc build on Linux: https://github.com/microsoft/LightGBM/pull/4274/checks?check_run_id=2550765412
• gcc build on Mac: https://github.com/microsoft/LightGBM/pull/4274/checks?check_run_id=2550765499
• MinGW build on Windows: https://github.com/microsoft/LightGBM/pull/4274/checks?check_run_id=2550765617

[StrikerRUS]

@jameslamb
Is this PR ready for review or supposed to be revised after the discussion with @david-cortes in #4273?

[jameslamb]

I think it can be reviewed. I interpreted the most recent comment in that discussion, #4273 (comment) , as saying "ok actually it's not related to use or missing use of extern C", so I can't think of other things to try.

Given the wide range of operating system + R version + compiler versions the R package is tested against, I'm fairly confident that this warning can be ignored and that it would improve the user experience to suppress it.

[StrikerRUS]

I've checked all CI logs to make sure that proposed if condition doesn't leave uncovered cases.
And I'm OK to ignore this warning.

[jameslamb]

I've checked all CI logs to make sure that proposed if condition doesn't leave uncovered cases.
And I'm OK to ignore this warning.

ok thanks very much!

[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.