[R-package] use safer pattern for error formatting (fixes #6212)

[jameslamb]

Fixes #6212.

Resolves warning -Wformat-security raised with R-devel as of a few days ago and clang-15:

lightgbm_R.cpp:159:3: warning: format string is not a string literal (potentially insecure) [-Wformat-security]

See #6212 (comment) for more details informing the changes in this PR.

Why not just suppress this warning?

CRAN has in the past not allowed compiler flags or pragmas in code that suppress warnings.

• LightGBM/build-cran-package.sh

  Lines 140 to 143 in 2ee3ec8

  	# Remove 'region', 'endregion', and 'warning' pragmas.
  	# This won't change the correctness of the code. CRAN does
  	# not allow you to use compiler flag '-Wno-unknown-pragmas' or
  	# pragmas that suppress warnings.

• [R-package] Suppress 'warning: unknown pragma' for CRAN package #3433

I couldn't find that stated explicitly in the CRAN policy (https://cran.r-project.org/web/packages/policies.html) or Writing R Extensions (https://cran.r-project.org/doc/manuals/R-exts.html), although there are some valid things there that note that warning-control flags can be compiler-specific and therefore add maintenance complexity to achieve portability. e.g. from https://cran.r-project.org/doc/manuals/R-exts.html#Portable-C-and-C_002b_002b-code

When writing a Makevars file for a package you intend to distribute, take care to ensure that it is not specific to your compiler: flags such as -O2 -Wall -pedantic (and all other -W flags: for the Oracle compilers these were used to pass arguments to compiler phases) are all specific to GCC (and compilers such as clang which aim to be options-compatible with it).

So unless there's some other performance, safety, or correctness issue I'm unaware of, I think the fix in this PR is a good approach to satisfying clang-15 and therefore the CRAN checks.

Notes for Reviewers

@shiyu1994 @guolinke @jmoralez this PR will unblock LightGBM's CI.

@david-cortes if you have other opinions on the approach I took here I'd welcome them.

[jameslamb]

@guolinke @shiyu1994 sorry for the @, but could one of you please review this? It should be non-controversial and quick, and it fixes an issue that's blocking other PRs from being merged.

Thank you.

[guolinke]

LGTM

[jameslamb]

thanks @guolinke !