Add xml_serialize()/xml_unserialize() tests for HTML-based documents …

…+ NEWS entry [#407]

NEWS.md

@@ -1,5 +1,8 @@
 # xml2 (development version)
 
+* `xml_serialize()` now includes also the document type so that `xml_unserialize()` works also for HTML documents (#407).
+
+
 # xml2 1.3.5
 
 * Small speedup for `xml_find_all()` (@mgirlich, #393).

tests/testthat/test-xml_serialize.R

@@ -2,6 +2,7 @@ x <- read_xml("<a>
   <b><c>123</c></b>
   <b><c>456</c></b>
   </a>")
+
 test_that("xml_serialize and xml_unserialize work with xml_document input", {
   out <- xml_unserialize(xml_serialize(x, NULL))
   expect_identical(as.character(x), as.character(out))
@@ -37,6 +38,20 @@ test_that("xml_serialize and xml_unserialize work with xml_nodeset input", {
   expect_identical(as.character(xml_unserialize(f)), as.character(b))
 })
 
+test_that("xml_serialize and xml_unserialize work with HTML-based xml_document input", {
+  file <- system.file("extdata", "r-project.html", package = "xml2")
+  x <- read_html(file)
+
+  out <- xml_unserialize(xml_serialize(x, NULL))
+  expect_identical(as.character(x), as.character(out))
+
+  f <- tempfile()
+  on.exit(unlink(f))
+
+  xml_serialize(x, f)
+  expect_identical(as.character(xml_unserialize(f)), as.character(x))
+})
+
 test_that("xml_unserialize throws an error if given a invalid object", {
   expect_error(xml_unserialize(serialize(1, NULL)), "Not a serialized xml2 object")
 })