fix: extend spell check to cover .qmd files

[d-morrison]

Fixes #69

The spelling R package only checks .R, .Rmd, and man pages by default, so .qmd files were never checked. This extends tests/spelling.R to also run spell_check_files() over all .qmd files using the existing WORDLIST.

Also fixes two typos now caught by the improved check: "oppenings" → "openings" and "anlyzing" → "analyzing".

Generated with Claude Code

[github-actions]

PR Preview Action v1.8.1-2-g6ad689f
Preview removed because the pull request was closed.
2026-06-05 00:53 UTC

[d-morrison]

@claude review

[claude]

Claude finished @d-morrison's task in 1m 33s —— View job

---

Code Review

• Gather context
• Read changed files
• Review code
• Post findings

---

Overall this is a clean, well-scoped PR. The two typo fixes are correct and the new .qmd spell-check logic is solid. A few minor observations below.

opportunities.qmd

Both fixes are correct:

• anlyzing → analyzing (line 19)
• oppenings → openings (line 66)

No issues.

tests/spelling.R

The new block is well-structured. Specific observations:

Good:

• Handles missing inst/WORDLIST gracefully with the if (file.exists(...)) guard.
• Handles the empty-file-list case with if (length(qmd_files) > 0).
• Uses lang = "en-US", which matches the Language: en-US field in DESCRIPTION — consistent with how spell_check_test() picks up the language.
• The comment explaining why spell_check_files() is needed (because spell_check_test() skips .qmd files) is a good use of a comment.

Minor: error-handling inconsistency

The existing check uses error = FALSE, which means spelling errors are printed but do not fail the test run:

spelling::spell_check_test(vignettes = TRUE, error = FALSE, skip_on_cran = TRUE)

The new .qmd check uses a hard stop(), which will fail the test run on any spelling error:

if (nrow(results) > 0) {
  print(results)
  stop("Spelling errors found in .qmd files")
}

If the intent is for both checks to behave the same way (fail CI on errors), the error = FALSE on the first call is likely a leftover from a project template and could be changed to error = TRUE for consistency. If error = FALSE is intentional (e.g., to allow currently-known errors in .Rmd/vignette files), the new .qmd behavior is intentionally stricter — which is fine since you've already fixed the errors caught by this check. Either way, a brief comment clarifying the intent would help future readers.

Fix this →

Nit: working-directory assumption

file.exists("inst/WORDLIST") and list.files(".", ...) are relative paths that assume tests are run from the project root. This is the standard convention for R packages and is already relied on by the existing code, so this is not a new concern — just worth being aware of.

---

Summary: The PR correctly solves the problem and the code is clean. The only actionable item is the error = FALSE / stop() inconsistency, which is worth addressing for clarity. Everything else is a nit or non-issue.

[d-morrison]

Claude Code CLI (local session) is working on this — paws off until I'm done.

[d-morrison]

Addressed the review:

• error = FALSE / stop() inconsistency: Added a comment documenting that the difference is intentional — the existing vignette/.Rmd check uses error = FALSE (report-only, so pre-existing/tolerated issues don't block CI), while the new .qmd check uses a hard stop() because .qmd spelling coverage is new and starts clean, so any new error should fail CI. Chose the comment route rather than flipping error = FALSE to TRUE, since changing the .Rmd/vignette behavior would expand this PR's scope and risk failing CI on untouched files.
• Working-directory assumption (nit): Acknowledged as non-actionable — the reviewer noted it's the standard R-package convention already relied on by existing code, not a new concern.

@claude review

[claude]

Claude finished @d-morrison's task in 26s —— View job

---

Code Review

• Read changed files
• Review addressed feedback
• Post findings

---

The previous review's actionable item has been handled well.

tests/spelling.R — addressed feedback

The added comment (lines 2–5) clearly and accurately documents the intentional difference between the two behaviors:

# error = FALSE: the vignette/.Rmd check reports spelling problems but does
# not fail the test run, so pre-existing/tolerated issues there don't block
# CI. The .qmd check below intentionally uses a hard stop() instead: .qmd
# spelling is newly covered and starts clean, so any new error should fail.

This is exactly the right call — changing error = FALSE to error = TRUE would have expanded the PR's scope and risked failing CI on unrelated files. The comment gives future readers everything they need to understand the asymmetry. No further changes needed.

---

This PR looks good to merge.