daniel1noble / BACE / 25331293731

91%

push

github

Fix R CMD check failures from previous commit

ca8265e passed devtools::test() locally + test-coverage on CI but
R-CMD-check failed across all 5 OS matrix entries. Three distinct
issues, all surfaced by the stricter R CMD check pipeline:

1. Unicode escapes ✓ ('✓') and ✗ ('✗') in R/convergence
   _plots.R triggered "conversion failure ... in 'mbcsToSbcs'" on
   locales without UTF-8 font fallback, which graphics::mtext
   propagates as an error -> tests aborted with 9 failures. Replaced
   the checkmark / cross characters with ASCII "[OK]" and "[X]"
   labels at six call sites in convergence_plots.R. Same semantics,
   plain-locale-safe.

2. tests/testthat/test-bace_options.R top-level called
   withr::local_options() but withr is not in DESCRIPTION's Suggests.
   R CMD check flagged "'::' or ':::' import not declared from:
   'withr'" (warning that --as-cran would treat as error). Removed
   the withr dependency entirely; per-test on.exit(reset_bace
   _options(), add = TRUE) handles state cleanup the same way.

3. .Rbuildignore did not exclude CLAUDE.md (project-level Claude
   instructions) or .claude/ (per-user settings). R CMD check noted
   "Non-standard file/directory found at top level: 'CLAUDE.md'".
   Added both to .Rbuildignore.

Verified locally: NOT_CRAN=true devtools::check() now reports
0 errors / 0 warnings / 0 notes.

Process note: in the future, run devtools::check() locally before
each push that touches R/ or tests/. devtools::test() exercises
the unit tests but skips the documentation / locale / namespace
audits that R CMD check (and the GHA R-CMD-check workflow) does.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

8 of 8 new or added lines in 1 file covered. (100.0%)

2476 of 3670 relevant lines covered (67.47%)

53.59 hits per line