// Copyright 2016 PDFium Authors. All rights reserved. // Use of this source code is governed by a BSD-style license that can be // found in the LICENSE file. #include "core/fxcrt/xml/cfx_xmlparser.h" #include #include "core/fxcrt/cfx_memorystream.h" #include "core/fxcrt/fx_codepage.h" #include "core/fxcrt/xml/cfx_xmlelement.h" #include "testing/gtest/include/gtest/gtest.h" #include "testing/test_support.h" #include "third_party/base/ptr_util.h" namespace { class CFX_XMLTestParser : public CFX_XMLParser { public: CFX_XMLTestParser(CFX_XMLNode* pParent, const RetainPtr& pStream) : CFX_XMLParser(pParent, pStream) {} ~CFX_XMLTestParser() override = default; FX_XmlSyntaxResult DoSyntaxParse() { return CFX_XMLParser::DoSyntaxParse(); } WideString GetTextData() const { return CFX_XMLParser::GetTextData(); } }; RetainPtr MakeProxy(const char* input) { auto stream = pdfium::MakeRetain( reinterpret_cast(const_cast(input)), strlen(input), false); return stream; } } // namespace TEST(CFX_XMLParserTest, CData) { const char* input = ""; const wchar_t* cdata = L"\n" L" if (a[1] < 3)\n" L" app.alert(\"Tclams\");\n" L" "; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::CData, parser.DoSyntaxParse()); ASSERT_EQ(cdata, parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, CDataWithInnerScript) { const char* input = "\n" " ]]>\n" ""; const wchar_t* cdata = L"\n" L" if (a[1] < 3)\n" L" app.alert(\"Tclams\");\n" L" \n" L" "; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::CData, parser.DoSyntaxParse()); ASSERT_EQ(cdata, parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, ArrowBangArrow) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, ArrowBangBracketArrow) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); // Parser walks to end of input. ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, IncompleteCData) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); // Parser walks to end of input. ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, UnClosedCData) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); // Parser walks to end of input. ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, EmptyCData) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::CData, parser.DoSyntaxParse()); ASSERT_EQ(L"", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, Comment) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, IncorrectCommentStart) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, CommentEmpty) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, CommentThreeDash) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, CommentTwoDash) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"\n ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, Entities) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L"BTjH\xab48&<>'\"", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, EntityOverflowHex) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L" ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, EntityOverflowDecimal) { const char* input = ""; auto stream = MakeProxy(input); auto root = pdfium::MakeUnique(L"ROOT"); CFX_XMLTestParser parser(root.get(), stream); ASSERT_EQ(FX_XmlSyntaxResult::ElementOpen, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::TagName, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriName, parser.DoSyntaxParse()); ASSERT_EQ(L"contentType", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::AttriValue, parser.DoSyntaxParse()); ASSERT_EQ(L"application/x-javascript", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementBreak, parser.DoSyntaxParse()); ASSERT_EQ(FX_XmlSyntaxResult::Text, parser.DoSyntaxParse()); ASSERT_EQ(L" ", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::ElementClose, parser.DoSyntaxParse()); ASSERT_EQ(L"script", parser.GetTextData()); ASSERT_EQ(FX_XmlSyntaxResult::EndOfString, parser.DoSyntaxParse()); } TEST(CFX_XMLParserTest, IsXMLNameChar) { EXPECT_FALSE(CFX_XMLTestParser::IsXMLNameChar(L'-', true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(L'-', false)); EXPECT_FALSE(CFX_XMLTestParser::IsXMLNameChar(0x2069, true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(0x2070, true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(0x2073, true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(0x218F, true)); EXPECT_FALSE(CFX_XMLTestParser::IsXMLNameChar(0x2190, true)); EXPECT_FALSE(CFX_XMLTestParser::IsXMLNameChar(0xFDEF, true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(0xFDF0, true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(0xFDF1, true)); EXPECT_TRUE(CFX_XMLTestParser::IsXMLNameChar(0xFFFD, true)); EXPECT_FALSE(CFX_XMLTestParser::IsXMLNameChar(0xFFFE, true)); }