Cover the Markdown tag filter's tab terminator and image descriptions

test/markdown/markdown_images_test.cc

@@ -83,3 +83,10 @@ TEST(image_title_with_escaped_quote) {
   EXPECT_EQ(result,
             "<p><img src=\"/l.png\" alt=\"logo\" title=\"it&#39;s\" /></p>\n");
 }
+
+TEST(image_description_spanning_a_soft_break) {
+  // The description becomes the alt attribute, which cannot hold a line break,
+  // so the break is flattened into a single space
+  const auto result{sourcemeta::core::markdown_to_html("![a\nb](/u)")};
+  EXPECT_EQ(result, "<p><img src=\"/u\" alt=\"a b\" /></p>\n");
+}

test/markdown/markdown_links_test.cc

@@ -655,3 +655,9 @@ TEST(link_inside_link_text_with_later_bracket_is_not_allowed) {
       sourcemeta::core::markdown_to_html("[a [b](/c) [d] e](/u)")};
   EXPECT_EQ(result, "<p>[a <a href=\"/c\">b</a> [d] e](/u)</p>\n");
 }
+
+TEST(reference_link_label_with_a_space_before_a_non_ascii_character) {
+  const auto result{sourcemeta::core::markdown_to_html(
+      "[click][a \u00e9]\n\n[a \u00e9]: https://example.com")};
+  EXPECT_EQ(result, "<p><a href=\"https://example.com\">click</a></p>\n");
+}

test/markdown/markdown_unsafe_test.cc

@@ -388,3 +388,12 @@ TEST(tagfilter_vertical_tab_does_not_end_tag_name) {
                     "<xmp\v>a\n"
                     "</div>\n");
 }
+
+TEST(tagfilter_tag_name_ended_by_a_tab) {
+  // The HTML tokenizer ends a tag name at a tab just as it does at a space, so
+  // the tag is still one of the disallowed ones and its opening angle bracket
+  // is escaped
+  const auto result{
+      sourcemeta::core::markdown_to_html("x <script\tsrc=y>", false)};
+  EXPECT_EQ(result, "<p>x &lt;script\tsrc=y></p>\n");
+}