read_html_keep_ansi()

R/markdown.R

@@ -62,7 +62,7 @@ markdown_body <- function(pkg, path, strip_header = FALSE) {
 markdown_path_html <- function(pkg, path, strip_header = FALSE) {
   html_path <- withr::local_tempfile()
   convert_markdown_to_html(pkg, path, html_path)
-  xml <- xml2::read_html(html_path, encoding = "UTF-8")
+  xml <- read_html_keep_ansi(html_path)
   if (!inherits(xml, "xml_node")) {
     return(NULL)
   }

R/tweak-page.R

@@ -96,14 +96,17 @@ tweak_useless_toc <- function(html) {
 
 # Update file on disk -----------------------------------------------------
 
-update_html <- function(path, tweak, ...) {
-
+read_html_keep_ansi <- function(path) {
   raw <- read_file(path)
   # Following the xml 1.0 spec, libxml2 drops low-bit ASCII characters
   # so we convert to \u2029, relying on downlit to convert back in
   # token_escape().
   raw <- gsub("\033", "\u2029", raw, fixed = TRUE)
-  html <- xml2::read_html(raw, encoding = "UTF-8")
+  xml2::read_html(charToRaw(raw), encoding = "UTF-8")
+}
+
+update_html <- function(path, tweak, ...) {
+  html <- read_html_keep_ansi(path)
   tweak(html, ...)
 
   xml2::write_html(html, path, format = FALSE)