Fix regression in read_html

jimhester · jimhester · commit 654385789de3 · 2020-04-04T15:09:20.000-04:00
Oh C string handling how I loathe you
diff --git a/NEWS.md b/NEWS.md
@@ -1,5 +1,7 @@
 # xml2 (development version)
 
+* `read_html()` now again works with HTML files with non-ASCII encodings (#293).
+
 # xml2 1.3.0
 
 * Removes the Rcpp dependency
diff --git a/src/xml2_doc.cpp b/src/xml2_doc.cpp
@@ -194,13 +194,13 @@ extern "C" SEXP doc_parse_file(
   if (as_html) {
     pDoc = htmlReadFile(
       path,
-      strncmp(encoding, "", 0) == 0 ? NULL : encoding,
+      encoding[0] == '\0' ? NULL : encoding,
       options
     );
   } else {
     pDoc = xmlReadFile(
       path,
-      strncmp(encoding, "", 0) == 0 ? NULL : encoding,
+      encoding[0] == '\0' ? NULL : encoding,
       options
     );
   }
diff --git a/tests/testthat/test-read-xml.R b/tests/testthat/test-read-xml.R
@@ -85,3 +85,15 @@ test_that("read_xml and read_html fail for bad status codes", {
     class = "http_404"
   )
 })
+
+test_that("read_html works with non-ASCII encodings", {
+  tmp <- tempfile()
+  on.exit(unlink(tmp))
+
+  writeLines("<html><body>\U2019</body></html>", tmp)
+  res <- read_html(tmp, encoding = "UTF-8")
+
+  expect_equal(as.character(res, options = ""),
+    "<!DOCTYPE html PUBLIC \"-//W3C//DTD HTML 4.0 Transitional//EN\" \"http://www.w3.org/TR/REC-html40/loose.dtd\">\n<html><body>\U2019</body></html>\n")
+})
+
