Add CDATA handling to parser, including a test for it.

google · Feb 17, 2015 · 58d5fad · 58d5fad
1 parent f9a515f
commit 58d5fad
Show file tree

Hide file tree

Showing 3 changed files with 30 additions and 9 deletions.
diff --git a/src/parser.c b/src/parser.c
@@ -793,7 +793,8 @@ static void maybe_flush_text_node_buffer(GumboParser* parser) {
   }
 
   assert(buffer_state->_type == GUMBO_NODE_WHITESPACE ||
-         buffer_state->_type == GUMBO_NODE_TEXT);
+         buffer_state->_type == GUMBO_NODE_TEXT ||
+         buffer_state->_type == GUMBO_NODE_CDATA);
   GumboNode* text_node = create_node(parser, buffer_state->_type);
   GumboText* text_node_data = &text_node->v.text;
   text_node_data->text = gumbo_string_buffer_to_string(
@@ -1019,7 +1020,8 @@ static GumboNode* insert_foreign_element(
 
 static void insert_text_token(GumboParser* parser, GumboToken* token) {
   assert(token->type == GUMBO_TOKEN_WHITESPACE ||
-         token->type == GUMBO_TOKEN_CHARACTER);
+         token->type == GUMBO_TOKEN_CHARACTER ||
+         token->type == GUMBO_TOKEN_CDATA);
   TextNodeBufferState* buffer_state = &parser->_parser_state->_text_node;
   if (buffer_state->_buffer.length == 0) {
     // Initialize position fields.
@@ -1030,6 +1032,8 @@ static void insert_text_token(GumboParser* parser, GumboToken* token) {
       parser, token->v.character, &buffer_state->_buffer);
   if (token->type == GUMBO_TOKEN_CHARACTER) {
     buffer_state->_type = GUMBO_NODE_TEXT;
+  } else if (token->type == GUMBO_TOKEN_CDATA) {
+    buffer_state->_type = GUMBO_NODE_CDATA;
   }
   gumbo_debug("Inserting text token '%c'.\n", token->v.character);
 }
@@ -2207,7 +2211,8 @@ static bool handle_in_body(GumboParser* parser, GumboToken* token) {
     reconstruct_active_formatting_elements(parser);
     insert_text_token(parser, token);
     return true;
-  } else if (token->type == GUMBO_TOKEN_CHARACTER) {
+  } else if (token->type == GUMBO_TOKEN_CHARACTER ||
+             token->type == GUMBO_TOKEN_CDATA) {
     reconstruct_active_formatting_elements(parser);
     insert_text_token(parser, token);
     set_frameset_not_ok(parser);
@@ -3492,6 +3497,7 @@ static bool handle_in_foreign_content(GumboParser* parser, GumboToken* token) {
     case GUMBO_TOKEN_WHITESPACE:
       insert_text_token(parser, token);
       return true;
+    case GUMBO_TOKEN_CDATA:
     case GUMBO_TOKEN_CHARACTER:
       insert_text_token(parser, token);
       set_frameset_not_ok(parser);

diff --git a/src/tokenizer.c b/src/tokenizer.c
@@ -319,7 +319,11 @@ static int ensure_lowercase(int c) {
   return c >= 'A' && c <= 'Z' ? c + 0x20 : c;
 }
 
-static GumboTokenType get_char_token_type(int c) {
+static GumboTokenType get_char_token_type(bool is_in_cdata, int c) {
+  if (is_in_cdata && c != -1) {
+    return GUMBO_TOKEN_CDATA;
+  }
+
   switch (c) {
     case '\t':
     case '\n':
@@ -479,11 +483,7 @@ static void finish_doctype_system_id(GumboParser* parser) {
 
 // Writes a single specified character to the output token.
 static void emit_char(GumboParser* parser, int c, GumboToken* output) {
-  if (parser->_tokenizer_state->_is_in_cdata) {
-    output->type = GUMBO_TOKEN_CDATA;
-  } else {
-    output->type = get_char_token_type(c);
-  }
+  output->type = get_char_token_type(parser->_tokenizer_state->_is_in_cdata, c);
   output->v.character = c;
   finish_token(parser, output);
 }

diff --git a/tests/parser.cc b/tests/parser.cc
@@ -1522,6 +1522,21 @@ TEST_F(GumboParserTest, ImplicitlyCloseLists) {
   ASSERT_EQ(1, GetChildCount(li2));
 }
 
+TEST_F(GumboParserTest, CData) {
+  Parse("<svg><![CDATA[this is text]]></svg>");
+
+  GumboNode* body;
+  GetAndAssertBody(root_, &body);
+  ASSERT_EQ(1, GetChildCount(body));
+
+  GumboNode* svg = GetChild(body, 0);
+  ASSERT_EQ(1, GetChildCount(svg));
+
+  GumboNode* cdata = GetChild(svg, 0);
+  ASSERT_EQ(GUMBO_NODE_CDATA, cdata->type);
+  EXPECT_STREQ("this is text", cdata->v.text.text);
+}
+
 TEST_F(GumboParserTest, FormattingTagsInHeading) {
   Parse("<h2>This is <b>old</h2>text");