Skip to content

urlchecker 2.0.0

Latest

Choose a tag to compare

@gaborcsardi gaborcsardi released this 09 Jul 06:07
  • url_check() now ignores the same HTTP status codes that CRAN ignores in
    its incoming URL checks: 202, 401, 403, and 429. This makes results match
    what CRAN reports and avoids false positives from servers that return.
    Set the _R_CHECK_URLS_HTTP_STATUS_IGNORE_REGEXP_ environment variable
    to override it.

  • url_check() now sends the same browser-like User-Agent header that
    CRAN uses for its URL checks. Set the _R_CHECK_URLS_CURL_USER_AGENT_
    environment variable to override it (#26).

  • url_check() no longer reports URLs that appear inside code chunks of
    .Rmd vignettes (#50).

  • url_check() now reads an optional .urlignore file (from the project
    root or from tools/). URLs matching its glob patterns are skipped and
    never requested, which is useful for links to private repositories or
    pages behind a login or captcha (#45).

  • The url_check() report no longer errors when a flagged URL is empty
    (e.g. a Markdown []() link); the empty URL is now reported without a
    source-line pointer (#47).

  • url_check() now checks URLs in BibTeX (.bib) bibliography files (#13).

  • url_check() now finds a GitHub token in the git credential store (via
    gitcreds, as configured by usethis or gh), in addition to the GITHUB_PAT
    environment variable. This lets checks of github.com URLs avoid the
    429: Too Many Requests errors that anonymous requests hit (#28).

  • url_check() gains a fail argument. When TRUE (the default) it throws
    an error after printing the report if any URLs are flagged, yielding a
    non-zero exit status for CI/CD workflows. Set fail = FALSE to return the
    results instead (#42).

  • The url_check() report now prints the file:line:col location of each
    flagged URL as a clickable hyperlink, so you can jump straight to the
    problem in supporting IDEs and terminals (#23).

  • url_check() now also works on non-package projects. path may be a
    directory that is not an R package (all supported files within are
    scanned), a single file, or a character vector mixing packages,
    directories and files. Supported file types are HTML, PDF, Rd, Markdown
    (.md, .markdown), R Markdown (.Rmd) and Quarto (.qmd).

  • url_check() now also checks URLs in Quarto (.qmd) vignettes,
    rendering them with quarto (the un-evaluated document, in the same way
    .Rmd vignettes are rendered with pandoc).

  • url_check() now errors if a VignetteBuilder package listed in the
    DESCRIPTION is not installed, as it is needed to render the vignettes
    for URL checking.

  • url_check() now errors, instead of silently skipping the file, when
    pandoc (for .md/.Rmd files) or quarto (for .qmd files) is needed to
    render a file but cannot be found on the PATH.

  • Handle URL fragments in redirects (#9).

  • url_check() now also handles package tarballs. This is useful to check
    the URL in the package vignettes.