xml2: Workaround bug for HTML documents (r-lib/xml2#407)

DESCRIPTION

@@ -1,5 +1,5 @@
 Package: marshal
-Version: 0.0.0-9033
+Version: 0.0.0-9034
 Title: Framework to Marshal Objects to be Used in Another R Process
 Depends:
     R (>= 3.2.0)

R/marshal.xml2.R

@@ -31,7 +31,10 @@ marshal.xml_nodeset <- function(xml2, ...) {
 
 marshal_xml2 <- function(xml2, ...) {
   res <- list(
-    marshalled = xml2::xml_serialize(xml2, connection = NULL)
+    marshalled = list(
+      doctype = xml2:::doc_type(xml2),
+      raw = xml2::xml_serialize(xml2, connection = NULL)
+    )
   )
   class(res) <- marshal_class(xml2)
 
@@ -47,7 +50,9 @@ marshal_xml2 <- function(xml2, ...) {
 
 unmarshal_xml2 <- function(xml2, ...) {
   object <- xml2[["marshalled"]]
-  res <- xml2::xml_unserialize(object)
+  args <- list(object[["raw"]])
+  if (object[["doctype"]] == "html") args$as_html <- TRUE
+  res <- do.call(xml2::xml_unserialize, args = args)
   stopifnot(identical(class(res), marshal_unclass(xml2)))
   res
 }

tests/marshal.xml2.R

@@ -13,4 +13,18 @@ if (requireNamespace("xml2", quietly = TRUE)) {
   doc2 <- unmarshal(doc_)
 
   stopifnot(all.equal(doc2, doc))
+
+  file <- system.file("extdata", "r-project.html", package = "xml2")
+  doc <- xml2::read_html(file)
+
+  ## Assert marshallability
+  stopifnot(marshallable(doc))
+
+  ## Marshal 'xml_document' object
+  doc_ <- marshal(doc)
+
+  ## Unmarshal 'xml_document' object
+  doc2 <- unmarshal(doc_)
+
+  stopifnot(all.equal(doc2, doc))
 }