Web API dependent HTML validation (#1495)

Author: katrinabrock

.github/workflows/R-CMD-check.yaml

@@ -66,6 +66,13 @@ jobs:
       actions: read
 
     steps:
+      - name: Start HTML validation server
+        run: |
+          docker run --rm -p 8888:8888 -d ghcr.io/validator/validator:latest &&
+          echo "W3C_MARKUP_VALIDATOR_BASEURL=http://0.0.0.0:8888" >> "$GITHUB_ENV"
+        if: runner.os == 'Linux'
+        shell: bash
+
       - uses: actions/checkout@v4
 
       - uses: r-lib/actions/setup-r@v2

DESCRIPTION

@@ -73,6 +73,7 @@ Suggests:
     downlit (>= 0.4.0),
     htmlwidgets,
     jsonlite,
+    curl,
     rstudioapi,
     miniUI,
     rsconnect (>= 0.4.3),

tests/rmd/split-section.Rmd

@@ -18,10 +18,16 @@ See chapter 2 now at \@ref(section-2)
 
 # Section 2
 
-## subsection 2 {#sub2}
+## subsection 21 {#sub2}
 
 ```{r iris-plot, fig.cap = "A plot"}
 plot(iris)
 ```
 
 See figure \@ref(fig:iris-plot)
+
+# subsection 22
+
+# Section 3
+
+## subsection 3

tests/test-rmd.R

@@ -1,3 +1,4 @@
+source('./testthat/helper-validate_html.R')
 # only run this when NOT_CRAN is true (e.g., on Travis CI)
 if (Sys.getenv('NOT_CRAN') == 'true') local({
   all_files = function() {
@@ -10,9 +11,16 @@ if (Sys.getenv('NOT_CRAN') == 'true') local({
     rmarkdown::render(f, envir = globalenv(), quiet = TRUE)
   }
 
+  validate_html(list.files("rmd", ".html$", full.names = TRUE))
+
   # split by section works correctly
   ## id is used for html file name
-  sections_files = c("section-1.html", "subsection-1.html", "section-2.html", "sub2.html")
+  sections_files = c(
+    "section-1.html", "subsection-1.html",
+    "section-2.html", "sub2.html", "subsection-22.html",
+    "section-3.html", "subsection-3.html"
+  )
+
   if (any(!file.exists(file.path("rmd", sections_files))))
     stop("Failed to generate sections files")
   ## reference is working correctly (see #787)

tests/testthat/helper-validate_html.R

@@ -0,0 +1,23 @@
+# the HTML validation server was started in GHA
+validate_html = function(files, server = Sys.getenv('W3C_MARKUP_VALIDATOR_BASEURL')) {
+  if (server == '') return()
+  res = lapply(files, function(f) {
+    h = curl::new_handle()
+    curl::handle_setform(handle = h, out = 'json', file = curl::form_file(f, 'text/html'))
+    jsonlite::fromJSON(rawToChar(curl::curl_fetch_memory(server, h)$content))
+  })
+  expected_errors = c(
+    'Attribute “number” not allowed on element “div” at this point.',
+    'CSS: “border-top”: “solid\\9” is not a “color” value.',
+    'CSS: “border-bottom”: “solid\\9” is not a “color” value.'
+  )
+  res = do.call(rbind, lapply(res, function(x) {
+    m <- x$messages$message[x$messages$type == 'error']
+    m <- setdiff(m, expected_errors)
+    if (length(m)) data.frame(file = x$url, messages = m)
+  }))
+  if (NROW(res) > 0) stop(
+    'HTML issues detected:\n',
+    paste0('  ', res$file, ': ', res$messages, collapse = '\n')
+  )
+}