R/qmd.R

Defines functions .quarto_html_for_CRAN urls_from_quarto_qmd_file url_db_from_package_qmd_vignettes

url_db_from_package_qmd_vignettes <- function(dir) {
  dir <- normalizePath(dir, winslash = "/")
  urls <- path <- character()
  docs <- Filter(file.exists, tools::pkgVignettes(dir = dir)$docs)
  # Only qmd vignettes are rendered with quarto; Rmd vignettes go through pandoc
  # in `url_db_from_package_rmd_vignettes()`.
  qfiles <- grep("[.]qmd$", docs, value = TRUE)
  for (qfile in qfiles) {
    qpath <- asNamespace("tools")$.file_path_relative_to_dir(
      normalizePath(qfile, winslash = "/"),
      dir
    )
    qurls <- urls_from_quarto_qmd_file(qfile)
    urls <- c(urls, qurls)
    path <- c(path, rep.int(qpath, length(qurls)))
  }
  tools$url_db(urls, path)
}

# Extract URLs from a `.qmd` file by rendering it to HTML with quarto and
# scraping the links. Errors if quarto is not available (it is required to
# render the file); returns `character()` if the render fails.
urls_from_quarto_qmd_file <- function(file) {
  if (!nzchar(Sys.which("quarto"))) {
    cli::cli_abort(c(
      "quarto is required to check URLs in {.file {basename(file)}}, \\
       but it was not found.",
      "i" = "Install quarto and make sure it is on the PATH."
    ))
  }
  # quarto writes supporting files (`*_files/`) next to the input, so render a
  # copy in a temp dir to keep the sources clean; the HTML itself is captured
  # from stdout (see `.quarto_html_for_CRAN()`).
  tdir <- tempfile()
  dir.create(tdir)
  on.exit(unlink(tdir, recursive = TRUE), add = TRUE)
  tin <- file.path(tdir, basename(file))
  file.copy(file, tin)
  out <- .quarto_html_for_CRAN(tin)
  if (!out$status && length(out$stdout)) {
    tfile <- tempfile(fileext = ".html")
    on.exit(unlink(tfile), add = TRUE)
    writeLines(out$stdout, tfile)
    tools$.get_urls_from_HTML_file(tfile)
  } else {
    character()
  }
}

# Render a `.qmd` to HTML (on stdout) for URL extraction only: `--no-execute`
# skips running code chunks (we only need the prose and links), and the `from`
# metadata adds pandoc's autolink_bare_uris extension so bare URLs are turned
# into links too, matching `.pandoc_md_for_CRAN2()` in rmd.R. `--output -`
# writes the HTML to stdout so we never place a file in the package sources.
.quarto_html_for_CRAN <- function(ifile) {
  asNamespace("tools")$.system_with_capture(
    "quarto",
    paste(
      "render",
      shQuote(normalizePath(ifile)),
      "--no-execute",
      "--to",
      "html",
      "--metadata",
      shQuote("from:markdown+autolink_bare_uris"),
      "--output",
      "-"
    )
  )
}

Try the urlchecker package in your browser

Any scripts or data that you put into this service are public.

urlchecker documentation built on July 8, 2026, 5:07 p.m.