Add xml:space="preserve" in XML output when appropriate

With current HTML escaping, sometimes we may produce an XML tag like

    <code_block>  </code_block>

Many XML parsers consider these spaces insignificant and strip them
out but we need this. There's actually a test case like this in
spec.txt (search "A code block can have all empty lines as its
content:") and without proper hinting, an external xml->html converter
will fail the spec.

XML standard covers this case. If xml:space is "preserve", then
whitespaces are significant and should be kept. Add this hint for
text, code, code_block, html_inline and html_block tags.

2 files changed, 29 insertions, 25 deletions

diff --git a/api_test/main.c b/api_test/main.c
index c30dc71..a843530 100644
--- a/api_test/main.c
+++ b/api_test/main.c
@@ -527,7 +527,9 @@ static void render_xml(test_batch_runner *runner) {
 
   static const char markdown[] = "foo *bar*\n"
                                  "\n"
-                                 "paragraph 2\n";
+                                 "paragraph 2\n"
+                                 "\n"
+                                 "```\ncode\n```\n";
   cmark_node *doc =
       cmark_parse_document(markdown, sizeof(markdown) - 1, CMARK_OPT_DEFAULT);
 
@@ -536,14 +538,16 @@ static void render_xml(test_batch_runner *runner) {
                       "<!DOCTYPE document SYSTEM \"CommonMark.dtd\">\n"
                       "<document xmlns=\"http://commonmark.org/xml/1.0\">\n"
                       "  <paragraph>\n"
-                      "    <text>foo </text>\n"
+                      "    <text xml:space=\"preserve\">foo </text>\n"
                       "    <emph>\n"
-                      "      <text>bar</text>\n"
+                      "      <text xml:space=\"preserve\">bar</text>\n"
                       "    </emph>\n"
                       "  </paragraph>\n"
                       "  <paragraph>\n"
-                      "    <text>paragraph 2</text>\n"
+                      "    <text xml:space=\"preserve\">paragraph 2</text>\n"
                       "  </paragraph>\n"
+                      "  <code_block xml:space=\"preserve\">code\n"
+                      "</code_block>\n"
                       "</document>\n",
          "render document");
   free(xml);
@@ -552,9 +556,9 @@ static void render_xml(test_batch_runner *runner) {
   STR_EQ(runner, xml, "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n"
                       "<!DOCTYPE document SYSTEM \"CommonMark.dtd\">\n"
                       "<paragraph sourcepos=\"1:1-1:9\">\n"
-                      "  <text sourcepos=\"1:1-1:4\">foo </text>\n"
+                      "  <text sourcepos=\"1:1-1:4\" xml:space=\"preserve\">foo </text>\n"
                       "  <emph sourcepos=\"1:5-1:9\">\n"
-                      "    <text sourcepos=\"1:6-1:8\">bar</text>\n"
+                      "    <text sourcepos=\"1:6-1:8\" xml:space=\"preserve\">bar</text>\n"
                       "  </emph>\n"
                       "</paragraph>\n",
          "render first paragraph with source pos");
@@ -902,41 +906,41 @@ static void source_pos(test_batch_runner *runner) {
                       "<!DOCTYPE document SYSTEM \"CommonMark.dtd\">\n"
                       "<document sourcepos=\"1:1-10:20\" xmlns=\"http://commonmark.org/xml/1.0\">\n"
                       "  <heading sourcepos=\"1:1-1:13\" level=\"1\">\n"
-                      "    <text sourcepos=\"1:3-1:5\">Hi </text>\n"
+                      "    <text sourcepos=\"1:3-1:5\" xml:space=\"preserve\">Hi </text>\n"
                       "    <emph sourcepos=\"1:6-1:12\">\n"
-                      "      <text sourcepos=\"1:7-1:11\">there</text>\n"
+                      "      <text sourcepos=\"1:7-1:11\" xml:space=\"preserve\">there</text>\n"
                       "    </emph>\n"
-                      "    <text sourcepos=\"1:13-1:13\">.</text>\n"
+                      "    <text sourcepos=\"1:13-1:13\" xml:space=\"preserve\">.</text>\n"
                       "  </heading>\n"
                       "  <paragraph sourcepos=\"3:1-4:42\">\n"
-                      "    <text sourcepos=\"3:1-3:14\">Hello “ </text>\n"
+                      "    <text sourcepos=\"3:1-3:14\" xml:space=\"preserve\">Hello “ </text>\n"
                       "    <link sourcepos=\"3:15-3:37\" destination=\"http://www.google.com\" title=\"\">\n"
-                      "      <text sourcepos=\"3:16-3:36\">http://www.google.com</text>\n"
+                      "      <text sourcepos=\"3:16-3:36\" xml:space=\"preserve\">http://www.google.com</text>\n"
                       "    </link>\n"
                       "    <softbreak />\n"
-                      "    <text sourcepos=\"4:1-4:6\">there </text>\n"
-                      "    <code sourcepos=\"4:8-4:9\">hi</code>\n"
-                      "    <text sourcepos=\"4:11-4:14\"> -- </text>\n"
+                      "    <text sourcepos=\"4:1-4:6\" xml:space=\"preserve\">there </text>\n"
+                      "    <code sourcepos=\"4:8-4:9\" xml:space=\"preserve\">hi</code>\n"
+                      "    <text sourcepos=\"4:11-4:14\" xml:space=\"preserve\"> -- </text>\n"
                       "    <link sourcepos=\"4:15-4:41\" destination=\"www.google.com\" title=\"ok\">\n"
-                      "      <text sourcepos=\"4:16-4:19\">okay</text>\n"
+                      "      <text sourcepos=\"4:16-4:19\" xml:space=\"preserve\">okay</text>\n"
                       "    </link>\n"
-                      "    <text sourcepos=\"4:42-4:42\">.</text>\n"
+                      "    <text sourcepos=\"4:42-4:42\" xml:space=\"preserve\">.</text>\n"
                       "  </paragraph>\n"
                       "  <block_quote sourcepos=\"6:1-10:20\">\n"
                       "    <list sourcepos=\"6:3-10:20\" type=\"ordered\" start=\"1\" delim=\"period\" tight=\"false\">\n"
                       "      <item sourcepos=\"6:3-8:1\">\n"
                       "        <paragraph sourcepos=\"6:6-7:10\">\n"
-                      "          <text sourcepos=\"6:6-6:10\">Okay.</text>\n"
+                      "          <text sourcepos=\"6:6-6:10\" xml:space=\"preserve\">Okay.</text>\n"
                       "          <softbreak />\n"
-                      "          <text sourcepos=\"7:6-7:10\">Sure.</text>\n"
+                      "          <text sourcepos=\"7:6-7:10\" xml:space=\"preserve\">Sure.</text>\n"
                       "        </paragraph>\n"
                       "      </item>\n"
                       "      <item sourcepos=\"9:3-10:20\">\n"
                       "        <paragraph sourcepos=\"9:6-10:20\">\n"
-                      "          <text sourcepos=\"9:6-9:15\">Yes, okay.</text>\n"
+                      "          <text sourcepos=\"9:6-9:15\" xml:space=\"preserve\">Yes, okay.</text>\n"
                       "          <softbreak />\n"
                       "          <image sourcepos=\"10:6-10:20\" destination=\"hi\" title=\"yes\">\n"
-                      "            <text sourcepos=\"10:8-10:9\">ok</text>\n"
+                      "            <text sourcepos=\"10:8-10:9\" xml:space=\"preserve\">ok</text>\n"
                       "          </image>\n"
                       "        </paragraph>\n"
                       "      </item>\n"
@@ -960,11 +964,11 @@ static void ref_source_pos(test_batch_runner *runner) {
                       "<!DOCTYPE document SYSTEM \"CommonMark.dtd\">\n"
                       "<document sourcepos=\"1:1-3:40\" xmlns=\"http://commonmark.org/xml/1.0\">\n"
                       "  <paragraph sourcepos=\"1:1-1:28\">\n"
-                      "    <text sourcepos=\"1:1-1:10\">Let's try </text>\n"
+                      "    <text sourcepos=\"1:1-1:10\" xml:space=\"preserve\">Let's try </text>\n"
                       "    <link sourcepos=\"1:11-1:21\" destination=\"https://github.com\" title=\"GitHub\">\n"
-                      "      <text sourcepos=\"1:12-1:20\">reference</text>\n"
+                      "      <text sourcepos=\"1:12-1:20\" xml:space=\"preserve\">reference</text>\n"
                       "    </link>\n"
-                      "    <text sourcepos=\"1:22-1:28\"> links.</text>\n"
+                      "    <text sourcepos=\"1:22-1:28\" xml:space=\"preserve\"> links.</text>\n"
                       "  </paragraph>\n"
                       "</document>\n",
          "sourcepos are as expected");
diff --git a/src/xml.c b/src/xml.c
index 4898cd2..48674cc 100644
--- a/src/xml.c
+++ b/src/xml.c
@@ -60,7 +60,7 @@ static int S_render_node(cmark_node *node, cmark_event_type ev_type,
     case CMARK_NODE_CODE:
     case CMARK_NODE_HTML_BLOCK:
     case CMARK_NODE_HTML_INLINE:
-      cmark_strbuf_puts(xml, ">");
+      cmark_strbuf_puts(xml, " xml:space=\"preserve\">");
       escape_xml(xml, node->as.literal.data, node->as.literal.len);
       cmark_strbuf_puts(xml, "</");
       cmark_strbuf_puts(xml, cmark_node_get_type_string(node));
@@ -100,7 +100,7 @@ static int S_render_node(cmark_node *node, cmark_event_type ev_type,
         escape_xml(xml, node->as.code.info.data, node->as.code.info.len);
         cmark_strbuf_putc(xml, '"');
       }
-      cmark_strbuf_puts(xml, ">");
+      cmark_strbuf_puts(xml, " xml:space=\"preserve\">");
       escape_xml(xml, node->as.code.literal.data, node->as.code.literal.len);
       cmark_strbuf_puts(xml, "</");
       cmark_strbuf_puts(xml, cmark_node_get_type_string(node));