From e1a41afbe146c9a976d96828a3a09a8a384741d9 Mon Sep 17 00:00:00 2001 From: Dan Sinclair Date: Thu, 12 Apr 2018 13:13:43 +0000 Subject: [xfa] Remove form checksum abilities This CL removes the code for verifying and creating checksums associated with form elements in XFA documents. This was the only code requiring the SAXReader which has also been removed. According to the XFA spec and application can decide which parts of the signatures are supported. This feature is being removed until we determine if/when it is needed. Bug: pdfium:1063 Change-Id: Iec2261282340f8fc72a1225d2e0d3e6ddf05edcb Reviewed-on: https://pdfium-review.googlesource.com/30150 Reviewed-by: Ryan Harrison Commit-Queue: dsinclair --- core/fxcrt/xml/cfx_saxcontext.cpp | 9 - core/fxcrt/xml/cfx_saxcontext.h | 25 - core/fxcrt/xml/cfx_saxreader.cpp | 744 ------------------------------ core/fxcrt/xml/cfx_saxreader.h | 174 ------- core/fxcrt/xml/cfx_saxreader_unittest.cpp | 152 ------ core/fxcrt/xml/cfx_saxreaderhandler.cpp | 128 ----- core/fxcrt/xml/cfx_saxreaderhandler.h | 48 -- core/fxcrt/xml/cfx_xmlparser.cpp | 30 +- core/fxcrt/xml/cfx_xmlparser.h | 6 - 9 files changed, 1 insertion(+), 1315 deletions(-) delete mode 100644 core/fxcrt/xml/cfx_saxcontext.cpp delete mode 100644 core/fxcrt/xml/cfx_saxcontext.h delete mode 100644 core/fxcrt/xml/cfx_saxreader.cpp delete mode 100644 core/fxcrt/xml/cfx_saxreader.h delete mode 100644 core/fxcrt/xml/cfx_saxreader_unittest.cpp delete mode 100644 core/fxcrt/xml/cfx_saxreaderhandler.cpp delete mode 100644 core/fxcrt/xml/cfx_saxreaderhandler.h (limited to 'core/fxcrt/xml') diff --git a/core/fxcrt/xml/cfx_saxcontext.cpp b/core/fxcrt/xml/cfx_saxcontext.cpp deleted file mode 100644 index 4e2f0c58c9..0000000000 --- a/core/fxcrt/xml/cfx_saxcontext.cpp +++ /dev/null @@ -1,9 +0,0 @@ -// Copyright 2017 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_saxcontext.h" - -CFX_SAXContext::CFX_SAXContext() : m_eNode(CFX_SAXItem::Type::Unknown) {} - -CFX_SAXContext::~CFX_SAXContext() {} diff --git a/core/fxcrt/xml/cfx_saxcontext.h b/core/fxcrt/xml/cfx_saxcontext.h deleted file mode 100644 index d4d74a385b..0000000000 --- a/core/fxcrt/xml/cfx_saxcontext.h +++ /dev/null @@ -1,25 +0,0 @@ -// Copyright 2017 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. - -// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com - -#ifndef CORE_FXCRT_XML_CFX_SAXCONTEXT_H_ -#define CORE_FXCRT_XML_CFX_SAXCONTEXT_H_ - -#include - -#include "core/fxcrt/fx_string.h" -#include "core/fxcrt/xml/cfx_saxreader.h" - -class CFX_SAXContext { - public: - CFX_SAXContext(); - ~CFX_SAXContext(); - - std::ostringstream m_TextBuf; - ByteString m_bsTagName; - CFX_SAXItem::Type m_eNode; -}; - -#endif // CORE_FXCRT_XML_CFX_SAXCONTEXT_H_ diff --git a/core/fxcrt/xml/cfx_saxreader.cpp b/core/fxcrt/xml/cfx_saxreader.cpp deleted file mode 100644 index 762144f4bb..0000000000 --- a/core/fxcrt/xml/cfx_saxreader.cpp +++ /dev/null @@ -1,744 +0,0 @@ -// Copyright 2014 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. - -// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com - -#include "core/fxcrt/xml/cfx_saxreader.h" - -#include -#include - -#include "core/fxcrt/fx_stream.h" -#include "core/fxcrt/xml/cfx_saxreaderhandler.h" -#include "third_party/base/ptr_util.h" -#include "third_party/base/stl_util.h" - -enum class CFX_SaxMode { - Text = 0, - NodeStart, - DeclOrComment, - DeclNode, - Comment, - CommentContent, - TagName, - TagAttributeName, - TagAttributeEqual, - TagAttributeValue, - TagMaybeClose, - TagClose, - TagEnd, - TargetData, -}; - -class CFX_SAXCommentContext { - public: - CFX_SAXCommentContext() : m_iHeaderCount(0), m_iTailCount(0) {} - int32_t m_iHeaderCount; - int32_t m_iTailCount; -}; - -namespace { - -const uint32_t kSaxFileBufSize = 32768; - -} // namespace - -CFX_SAXFile::CFX_SAXFile() - : m_dwStart(0), - m_dwEnd(0), - m_dwCur(0), - m_pBuf(nullptr), - m_dwBufSize(0), - m_dwBufIndex(0) {} - -CFX_SAXFile::~CFX_SAXFile() {} - -bool CFX_SAXFile::StartFile(const RetainPtr& pFile, - uint32_t dwStart, - uint32_t dwLen) { - ASSERT(!m_pFile && pFile); - uint32_t dwSize = pFile->GetSize(); - if (dwStart >= dwSize) - return false; - - if (dwLen == static_cast(-1) || dwStart + dwLen > dwSize) - dwLen = dwSize - dwStart; - - if (dwLen == 0) - return false; - - m_dwBufSize = std::min(dwLen, kSaxFileBufSize); - m_pBuf = FX_Alloc(uint8_t, m_dwBufSize); - if (!pFile->ReadBlock(m_pBuf, dwStart, m_dwBufSize)) - return false; - - m_dwStart = dwStart; - m_dwEnd = dwStart + dwLen; - m_dwCur = dwStart; - m_pFile = pFile; - m_dwBufIndex = 0; - return true; -} - -bool CFX_SAXFile::ReadNextBlock() { - ASSERT(m_pFile); - uint32_t dwSize = m_dwEnd - m_dwCur; - if (dwSize == 0) { - return false; - } - m_dwBufSize = std::min(dwSize, kSaxFileBufSize); - if (!m_pFile->ReadBlock(m_pBuf, m_dwCur, m_dwBufSize)) { - return false; - } - m_dwBufIndex = 0; - return true; -} - -void CFX_SAXFile::Reset() { - if (m_pBuf) { - FX_Free(m_pBuf); - m_pBuf = nullptr; - } - m_pFile = nullptr; -} - -CFX_SAXReader::CFX_SAXReader() - : m_File(), - m_pHandler(nullptr), - m_iState(-1), - m_dwItemID(0), - m_dwParseMode(0) { - m_Data.reserve(256); - m_Name.reserve(256); -} - -CFX_SAXReader::~CFX_SAXReader() { - Reset(); -} - -void CFX_SAXReader::Reset() { - m_File.Reset(); - m_iState = -1; - m_Stack = std::stack>(); - m_dwItemID = 0; - m_SkipStack = std::stack(); - m_SkipChar = 0; - m_pCommentContext.reset(); - ClearData(); - ClearName(); -} - -void CFX_SAXReader::Push() { - std::unique_ptr pNew = - pdfium::MakeUnique(++m_dwItemID); - if (!m_Stack.empty()) - pNew->m_bSkip = m_Stack.top()->m_bSkip; - m_Stack.push(std::move(pNew)); -} - -void CFX_SAXReader::Pop() { - if (!m_Stack.empty()) - m_Stack.pop(); -} - -CFX_SAXItem* CFX_SAXReader::GetCurrentItem() const { - return m_Stack.empty() ? nullptr : m_Stack.top().get(); -} - -void CFX_SAXReader::ClearData() { - m_Data.clear(); - m_iEntityStart = -1; -} - -void CFX_SAXReader::ClearName() { - m_Name.clear(); -} - -void CFX_SAXReader::AppendToData(uint8_t ch) { - m_Data.push_back(ch); -} - -void CFX_SAXReader::AppendToName(uint8_t ch) { - m_Name.push_back(ch); -} - -void CFX_SAXReader::BackUpAndReplaceDataAt(int32_t index, uint8_t ch) { - ASSERT(index > -1); - m_Data.erase(m_Data.begin() + index, m_Data.end()); - AppendToData(ch); -} - -int32_t CFX_SAXReader::CurrentDataIndex() const { - return pdfium::CollectionSize(m_Data) - 1; -} - -bool CFX_SAXReader::IsEntityStart(uint8_t ch) const { - return m_iEntityStart == -1 && ch == '&'; -} - -bool CFX_SAXReader::IsEntityEnd(uint8_t ch) const { - return m_iEntityStart != -1 && ch == ';'; -} - -bool CFX_SAXReader::SkipSpace(uint8_t ch) { - return (m_dwParseMode & CFX_SaxParseMode_NotSkipSpace) == 0 && ch < 0x21; -} - -int32_t CFX_SAXReader::StartParse( - const RetainPtr& pFile, - uint32_t dwStart, - uint32_t dwLen, - uint32_t dwParseMode) { - Reset(); - if (!m_File.StartFile(pFile, dwStart, dwLen)) - return -1; - - m_iState = 0; - m_eMode = CFX_SaxMode::Text; - m_ePrevMode = CFX_SaxMode::Text; - m_bCharData = false; - m_dwDataOffset = 0; - m_dwParseMode = dwParseMode; - m_Stack.push(pdfium::MakeUnique(++m_dwItemID)); - return 0; -} - -int32_t CFX_SAXReader::ContinueParse() { - if (m_iState < 0 || m_iState > 99) - return m_iState; - - while (m_File.m_dwCur < m_File.m_dwEnd) { - uint32_t& index = m_File.m_dwBufIndex; - uint32_t size = m_File.m_dwBufSize; - const uint8_t* pBuf = m_File.m_pBuf; - while (index < size) { - m_CurByte = pBuf[index]; - ParseInternal(); - index++; - } - m_File.m_dwCur += index; - m_iState = (m_File.m_dwCur - m_File.m_dwStart) * 100 / - (m_File.m_dwEnd - m_File.m_dwStart); - if (m_File.m_dwCur >= m_File.m_dwEnd) - break; - if (!m_File.ReadNextBlock()) { - m_iState = -2; - break; - } - m_dwDataOffset = 0; - } - return m_iState; -} - -void CFX_SAXReader::ParseInternal() { - switch (m_eMode) { - case CFX_SaxMode::Text: - ParseText(); - break; - case CFX_SaxMode::NodeStart: - ParseNodeStart(); - break; - case CFX_SaxMode::DeclOrComment: - ParseDeclOrComment(); - break; - case CFX_SaxMode::DeclNode: - ParseDeclNode(); - break; - case CFX_SaxMode::Comment: - ParseComment(); - break; - case CFX_SaxMode::CommentContent: - ParseCommentContent(); - break; - case CFX_SaxMode::TagName: - ParseTagName(); - break; - case CFX_SaxMode::TagAttributeName: - ParseTagAttributeName(); - break; - case CFX_SaxMode::TagAttributeEqual: - ParseTagAttributeEqual(); - break; - case CFX_SaxMode::TagAttributeValue: - ParseTagAttributeValue(); - break; - case CFX_SaxMode::TagMaybeClose: - ParseMaybeClose(); - break; - case CFX_SaxMode::TagClose: - ParseTagClose(); - break; - case CFX_SaxMode::TagEnd: - ParseTagEnd(); - break; - case CFX_SaxMode::TargetData: - ParseTargetData(); - break; - } -} - -void CFX_SAXReader::ParseChar(uint8_t ch) { - AppendToData(ch); - if (IsEntityStart(ch)) { - m_iEntityStart = CurrentDataIndex(); - return; - } - if (!IsEntityEnd(ch)) - return; - - // No matter what, we're no longer in an entity. - ASSERT(m_iEntityStart > -1); - int32_t iSaveStart = m_iEntityStart; - m_iEntityStart = -1; - - // NOTE: Relies on negative lengths being treated as empty strings. - ByteString csEntity(m_Data.data() + iSaveStart + 1, - CurrentDataIndex() - iSaveStart - 1); - int32_t iLen = csEntity.GetLength(); - if (iLen == 0) - return; - - if (csEntity[0] == '#') { - if ((m_dwParseMode & CFX_SaxParseMode_NotConvert_sharp) == 0) { - ch = 0; - uint8_t w; - if (iLen > 1 && csEntity[1] == 'x') { - for (int32_t i = 2; i < iLen; i++) { - w = csEntity[i]; - if (w >= '0' && w <= '9') - ch = (ch << 4) + w - '0'; - else if (w >= 'A' && w <= 'F') - ch = (ch << 4) + w - 55; - else if (w >= 'a' && w <= 'f') - ch = (ch << 4) + w - 87; - else - break; - } - } else { - for (int32_t i = 1; i < iLen; i++) { - w = csEntity[i]; - if (w < '0' || w > '9') - break; - ch = ch * 10 + w - '0'; - } - } - if (ch != 0) - BackUpAndReplaceDataAt(iSaveStart, ch); - } - return; - } - if (csEntity == "amp") { - if ((m_dwParseMode & CFX_SaxParseMode_NotConvert_amp) == 0) - BackUpAndReplaceDataAt(iSaveStart, '&'); - return; - } - if (csEntity == "lt") { - if ((m_dwParseMode & CFX_SaxParseMode_NotConvert_lt) == 0) - BackUpAndReplaceDataAt(iSaveStart, '<'); - return; - } - if (csEntity == "gt") { - if ((m_dwParseMode & CFX_SaxParseMode_NotConvert_gt) == 0) - BackUpAndReplaceDataAt(iSaveStart, '>'); - return; - } - if (csEntity == "apos") { - if ((m_dwParseMode & CFX_SaxParseMode_NotConvert_apos) == 0) - BackUpAndReplaceDataAt(iSaveStart, '\''); - return; - } - if (csEntity == "quot") { - if ((m_dwParseMode & CFX_SaxParseMode_NotConvert_quot) == 0) - BackUpAndReplaceDataAt(iSaveStart, '\"'); - return; - } -} - -void CFX_SAXReader::ParseText() { - if (m_CurByte == '<') { - if (!m_Data.empty()) { - NotifyData(); - ClearData(); - } - Push(); - m_dwNodePos = m_File.m_dwCur + m_File.m_dwBufIndex; - m_eMode = CFX_SaxMode::NodeStart; - return; - } - if (m_Data.empty() && SkipSpace(m_CurByte)) - return; - - ParseChar(m_CurByte); -} - -void CFX_SAXReader::ParseNodeStart() { - if (m_CurByte == '?') { - GetCurrentItem()->m_eNode = CFX_SAXItem::Type::Instruction; - m_eMode = CFX_SaxMode::TagName; - return; - } - if (m_CurByte == '!') { - m_eMode = CFX_SaxMode::DeclOrComment; - return; - } - if (m_CurByte == '/') { - m_eMode = CFX_SaxMode::TagEnd; - return; - } - if (m_CurByte == '>') { - Pop(); - m_eMode = CFX_SaxMode::Text; - return; - } - if (m_CurByte > 0x20) { - m_dwDataOffset = m_File.m_dwBufIndex; - GetCurrentItem()->m_eNode = CFX_SAXItem::Type::Tag; - m_eMode = CFX_SaxMode::TagName; - AppendToData(m_CurByte); - } -} - -void CFX_SAXReader::ParseDeclOrComment() { - if (m_CurByte == '-') { - m_eMode = CFX_SaxMode::Comment; - GetCurrentItem()->m_eNode = CFX_SAXItem::Type::Comment; - if (!m_pCommentContext) - m_pCommentContext = pdfium::MakeUnique(); - m_pCommentContext->m_iHeaderCount = 1; - m_pCommentContext->m_iTailCount = 0; - return; - } - m_eMode = CFX_SaxMode::DeclNode; - m_dwDataOffset = m_File.m_dwBufIndex; - m_SkipChar = '>'; - m_SkipStack.push('>'); - SkipNode(); -} - -void CFX_SAXReader::ParseComment() { - m_pCommentContext->m_iHeaderCount = 2; - m_dwNodePos = m_File.m_dwCur + m_File.m_dwBufIndex; - m_eMode = CFX_SaxMode::CommentContent; -} - -void CFX_SAXReader::ParseCommentContent() { - if (m_CurByte == '-') { - m_pCommentContext->m_iTailCount++; - return; - } - if (m_CurByte == '>' && m_pCommentContext->m_iTailCount == 2) { - NotifyTargetData(); - ClearData(); - Pop(); - m_eMode = CFX_SaxMode::Text; - return; - } - while (m_pCommentContext->m_iTailCount > 0) { - AppendToData('-'); - m_pCommentContext->m_iTailCount--; - } - AppendToData(m_CurByte); -} - -void CFX_SAXReader::ParseDeclNode() { - SkipNode(); -} - -void CFX_SAXReader::ParseTagName() { - if (m_CurByte < 0x21 || m_CurByte == '/' || m_CurByte == '>' || - m_CurByte == '?') { - NotifyEnter(); - ClearData(); - if (m_CurByte < 0x21) { - ClearName(); - m_eMode = CFX_SaxMode::TagAttributeName; - } else if (m_CurByte == '/' || m_CurByte == '?') { - m_ePrevMode = m_eMode; - m_eMode = CFX_SaxMode::TagMaybeClose; - } else { - NotifyBreak(); - m_eMode = CFX_SaxMode::Text; - } - } else { - AppendToData(m_CurByte); - } -} - -void CFX_SAXReader::ParseTagAttributeName() { - if (m_CurByte < 0x21 || m_CurByte == '=') { - if (m_Name.empty() && m_CurByte < 0x21) - return; - - m_SkipChar = 0; - m_eMode = m_CurByte == '=' ? CFX_SaxMode::TagAttributeValue - : CFX_SaxMode::TagAttributeEqual; - ClearData(); - return; - } - if (m_CurByte == '/' || m_CurByte == '>' || m_CurByte == '?') { - if (m_CurByte == '/' || m_CurByte == '?') { - m_ePrevMode = m_eMode; - m_eMode = CFX_SaxMode::TagMaybeClose; - } else { - NotifyBreak(); - m_eMode = CFX_SaxMode::Text; - } - return; - } - if (m_Name.empty()) - m_dwDataOffset = m_File.m_dwBufIndex; - AppendToName(m_CurByte); -} - -void CFX_SAXReader::ParseTagAttributeEqual() { - if (m_CurByte == '=') { - m_SkipChar = 0; - m_eMode = CFX_SaxMode::TagAttributeValue; - return; - } - if (GetCurrentItem()->m_eNode == CFX_SAXItem::Type::Instruction) { - AppendToName(0x20); - m_eMode = CFX_SaxMode::TargetData; - ParseTargetData(); - } -} - -void CFX_SAXReader::ParseTagAttributeValue() { - if (m_SkipChar) { - if (m_SkipChar == m_CurByte) { - NotifyAttribute(); - ClearData(); - ClearName(); - m_SkipChar = 0; - m_eMode = CFX_SaxMode::TagAttributeName; - return; - } - ParseChar(m_CurByte); - return; - } - if (m_CurByte < 0x21) { - return; - } - if (m_Data.empty()) { - if (m_CurByte == '\'' || m_CurByte == '\"') - m_SkipChar = m_CurByte; - } -} - -void CFX_SAXReader::ParseMaybeClose() { - if (m_CurByte == '>') { - if (GetCurrentItem()->m_eNode == CFX_SAXItem::Type::Instruction) { - NotifyTargetData(); - ClearData(); - ClearName(); - } - ParseTagClose(); - m_eMode = CFX_SaxMode::Text; - } else if (m_ePrevMode == CFX_SaxMode::TagName) { - AppendToData('/'); - m_eMode = CFX_SaxMode::TagName; - m_ePrevMode = CFX_SaxMode::Text; - ParseTagName(); - } else if (m_ePrevMode == CFX_SaxMode::TagAttributeName) { - AppendToName('/'); - m_eMode = CFX_SaxMode::TagAttributeName; - m_ePrevMode = CFX_SaxMode::Text; - ParseTagAttributeName(); - } else if (m_ePrevMode == CFX_SaxMode::TargetData) { - AppendToName('?'); - m_eMode = CFX_SaxMode::TargetData; - m_ePrevMode = CFX_SaxMode::Text; - ParseTargetData(); - } -} -void CFX_SAXReader::ParseTagClose() { - m_dwNodePos = m_File.m_dwCur + m_File.m_dwBufIndex; - NotifyClose(); - Pop(); -} -void CFX_SAXReader::ParseTagEnd() { - if (m_CurByte < 0x21) { - return; - } - if (m_CurByte == '>') { - Pop(); - m_dwNodePos = m_File.m_dwCur + m_File.m_dwBufIndex; - NotifyEnd(); - ClearData(); - Pop(); - m_eMode = CFX_SaxMode::Text; - } else { - ParseChar(m_CurByte); - } -} -void CFX_SAXReader::ParseTargetData() { - if (m_CurByte == '?') { - m_ePrevMode = m_eMode; - m_eMode = CFX_SaxMode::TagMaybeClose; - } else { - AppendToName(m_CurByte); - } -} -void CFX_SAXReader::SkipNode() { - if (m_SkipChar == '\'' || m_SkipChar == '\"') { - if (m_CurByte != m_SkipChar) - return; - - ASSERT(!m_SkipStack.empty()); - m_SkipStack.pop(); - m_SkipChar = !m_SkipStack.empty() ? m_SkipStack.top() : 0; - return; - } - switch (m_CurByte) { - case '<': - m_SkipChar = '>'; - m_SkipStack.push('>'); - break; - case '[': - m_SkipChar = ']'; - m_SkipStack.push(']'); - break; - case '(': - m_SkipChar = ')'; - m_SkipStack.push(')'); - break; - case '\'': - m_SkipChar = '\''; - m_SkipStack.push('\''); - break; - case '\"': - m_SkipChar = '\"'; - m_SkipStack.push('\"'); - break; - default: - if (m_CurByte == m_SkipChar) { - m_SkipStack.pop(); - m_SkipChar = !m_SkipStack.empty() ? m_SkipStack.top() : 0; - if (m_SkipStack.empty() && m_CurByte == '>') { - if (m_Data.size() >= 9 && memcmp(m_Data.data(), "[CDATA[", 7) == 0 && - memcmp(m_Data.data() + m_Data.size() - 2, "]]", 2) == 0) { - Pop(); - m_Data.erase(m_Data.begin(), m_Data.begin() + 7); - m_Data.erase(m_Data.end() - 2, m_Data.end()); - m_bCharData = true; - NotifyData(); - m_bCharData = false; - } else { - Pop(); - } - ClearData(); - m_eMode = CFX_SaxMode::Text; - } - } - break; - } - if (!m_SkipStack.empty()) - ParseChar(m_CurByte); -} - -void CFX_SAXReader::NotifyData() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Tag) - m_pHandler->OnTagData( - pItem->m_pNode, - m_bCharData ? CFX_SAXItem::Type::CharData : CFX_SAXItem::Type::Text, - ByteStringView(m_Data), m_File.m_dwCur + m_dwDataOffset); -} - -void CFX_SAXReader::NotifyEnter() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Tag || - pItem->m_eNode == CFX_SAXItem::Type::Instruction) { - pItem->m_pNode = m_pHandler->OnTagEnter(ByteStringView(m_Data), - pItem->m_eNode, m_dwNodePos); - } -} - -void CFX_SAXReader::NotifyAttribute() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Tag || - pItem->m_eNode == CFX_SAXItem::Type::Instruction) { - m_pHandler->OnTagAttribute(pItem->m_pNode, ByteStringView(m_Name), - ByteStringView(m_Data)); - } -} - -void CFX_SAXReader::NotifyBreak() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Tag) - m_pHandler->OnTagBreak(pItem->m_pNode); -} - -void CFX_SAXReader::NotifyClose() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Tag || - pItem->m_eNode == CFX_SAXItem::Type::Instruction) { - m_pHandler->OnTagClose(pItem->m_pNode, m_dwNodePos); - } -} - -void CFX_SAXReader::NotifyEnd() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Tag) - m_pHandler->OnTagEnd(pItem->m_pNode, ByteStringView(m_Data), m_dwNodePos); -} - -void CFX_SAXReader::NotifyTargetData() { - if (!m_pHandler) - return; - - CFX_SAXItem* pItem = GetCurrentItem(); - if (!pItem) - return; - - if (pItem->m_eNode == CFX_SAXItem::Type::Instruction) { - m_pHandler->OnTargetData(pItem->m_pNode, pItem->m_eNode, - ByteStringView(m_Name), m_dwNodePos); - } else if (pItem->m_eNode == CFX_SAXItem::Type::Comment) { - m_pHandler->OnTargetData(pItem->m_pNode, pItem->m_eNode, - ByteStringView(m_Data), m_dwNodePos); - } -} - -void CFX_SAXReader::SkipCurrentNode() { - CFX_SAXItem* pItem = GetCurrentItem(); - if (pItem) - pItem->m_bSkip = true; -} diff --git a/core/fxcrt/xml/cfx_saxreader.h b/core/fxcrt/xml/cfx_saxreader.h deleted file mode 100644 index 9ff755d053..0000000000 --- a/core/fxcrt/xml/cfx_saxreader.h +++ /dev/null @@ -1,174 +0,0 @@ -// Copyright 2014 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. - -// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com - -#ifndef CORE_FXCRT_XML_CFX_SAXREADER_H_ -#define CORE_FXCRT_XML_CFX_SAXREADER_H_ - -#include -#include -#include - -#include "core/fxcrt/fx_string.h" -#include "core/fxcrt/retain_ptr.h" - -class CFX_SAXCommentContext; -class CFX_SAXContext; -class IFX_SeekableReadStream; -enum class CFX_SaxMode; - -class CFX_SAXItem { - public: - enum class Type { - Unknown = 0, - Instruction, - Declaration, - Comment, - Tag, - Text, - CharData, - }; - - explicit CFX_SAXItem(uint32_t id) - : m_pNode(nullptr), m_eNode(Type::Unknown), m_dwID(id), m_bSkip(false) {} - - CFX_SAXContext* m_pNode; - Type m_eNode; - const uint32_t m_dwID; - bool m_bSkip; -}; - -class CFX_SAXFile { - public: - CFX_SAXFile(); - ~CFX_SAXFile(); - - bool StartFile(const RetainPtr& pFile, - uint32_t dwStart, - uint32_t dwLen); - bool ReadNextBlock(); - void Reset(); - - RetainPtr m_pFile; - uint32_t m_dwStart; - uint32_t m_dwEnd; - uint32_t m_dwCur; - uint8_t* m_pBuf; - uint32_t m_dwBufSize; - uint32_t m_dwBufIndex; -}; - -enum CFX_SaxParseMode { - CFX_SaxParseMode_NotConvert_amp = 1 << 0, - CFX_SaxParseMode_NotConvert_lt = 1 << 1, - CFX_SaxParseMode_NotConvert_gt = 1 << 2, - CFX_SaxParseMode_NotConvert_apos = 1 << 3, - CFX_SaxParseMode_NotConvert_quot = 1 << 4, - CFX_SaxParseMode_NotConvert_sharp = 1 << 5, - CFX_SaxParseMode_NotSkipSpace = 1 << 6 -}; - -class CFX_SAXReader { - public: - class HandlerIface { - public: - virtual ~HandlerIface() {} - virtual CFX_SAXContext* OnTagEnter(const ByteStringView& bsTagName, - CFX_SAXItem::Type eType, - uint32_t dwStartPos) = 0; - virtual void OnTagAttribute(CFX_SAXContext* pTag, - const ByteStringView& bsAttri, - const ByteStringView& bsValue) = 0; - virtual void OnTagBreak(CFX_SAXContext* pTag) = 0; - virtual void OnTagData(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos) = 0; - virtual void OnTagClose(CFX_SAXContext* pTag, uint32_t dwEndPos) = 0; - virtual void OnTagEnd(CFX_SAXContext* pTag, - const ByteStringView& bsTagName, - uint32_t dwEndPos) = 0; - virtual void OnTargetData(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos) = 0; - }; - - CFX_SAXReader(); - ~CFX_SAXReader(); - - int32_t StartParse(const RetainPtr& pFile, - uint32_t dwStart = 0, - uint32_t dwLen = -1, - uint32_t dwParseMode = 0); - int32_t ContinueParse(); - void SetHandler(HandlerIface* pHandler) { m_pHandler = pHandler; } - - private: - void ParseInternal(); - void SkipCurrentNode(); - void AppendData(uint8_t ch); - void AppendName(uint8_t ch); - void ParseText(); - void ParseNodeStart(); - void ParseInstruction(); - void ParseDeclOrComment(); - void ParseDeclNode(); - void ParseComment(); - void ParseCommentContent(); - void ParseTagName(); - void ParseTagAttributeName(); - void ParseTagAttributeEqual(); - void ParseTagAttributeValue(); - void ParseMaybeClose(); - void ParseTagClose(); - void ParseTagEnd(); - void ParseTargetData(); - void Reset(); - void ClearData(); - void ClearName(); - void AppendToData(uint8_t ch); - void AppendToName(uint8_t ch); - void BackUpAndReplaceDataAt(int32_t index, uint8_t ch); - bool IsEntityStart(uint8_t ch) const; - bool IsEntityEnd(uint8_t ch) const; - int32_t CurrentDataIndex() const; - void Push(); - void Pop(); - CFX_SAXItem* GetCurrentItem() const; - bool SkipSpace(uint8_t ch); - void SkipNode(); - void NotifyData(); - void NotifyEnter(); - void NotifyAttribute(); - void NotifyBreak(); - void NotifyClose(); - void NotifyEnd(); - void NotifyTargetData(); - void ReallocDataBuffer(); - void ReallocNameBuffer(); - void ParseChar(uint8_t ch); - - CFX_SAXFile m_File; - HandlerIface* m_pHandler; - int32_t m_iState; - std::stack> m_Stack; - uint32_t m_dwItemID; - CFX_SaxMode m_eMode; - CFX_SaxMode m_ePrevMode; - bool m_bCharData; - uint8_t m_CurByte; - uint32_t m_dwDataOffset; - std::stack m_SkipStack; - uint8_t m_SkipChar; - uint32_t m_dwNodePos; - std::vector m_Data; - int32_t m_iEntityStart; // Index into m_Data. - std::vector m_Name; - uint32_t m_dwParseMode; - std::unique_ptr m_pCommentContext; -}; - -#endif // CORE_FXCRT_XML_CFX_SAXREADER_H_ diff --git a/core/fxcrt/xml/cfx_saxreader_unittest.cpp b/core/fxcrt/xml/cfx_saxreader_unittest.cpp deleted file mode 100644 index 13d06325a7..0000000000 --- a/core/fxcrt/xml/cfx_saxreader_unittest.cpp +++ /dev/null @@ -1,152 +0,0 @@ -// Copyright 2017 The 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_saxreader.h" -#include "core/fxcrt/cfx_memorystream.h" -#include "testing/gmock/include/gmock/gmock.h" -#include "testing/gtest/include/gtest/gtest.h" -#include "testing/test_support.h" - -using testing::_; -using testing::Eq; -using testing::Return; - -namespace { - -class MockHandler : public CFX_SAXReader::HandlerIface { - public: - MOCK_METHOD3(OnTagEnter, - CFX_SAXContext*(const ByteStringView& bsTagName, - CFX_SAXItem::Type eType, - uint32_t dwStartPos)); - MOCK_METHOD3(OnTagAttribute, - void(CFX_SAXContext* pTag, - const ByteStringView& bsAttri, - const ByteStringView& bsValue)); - MOCK_METHOD1(OnTagBreak, void(CFX_SAXContext* pTag)); - MOCK_METHOD4(OnTagData, - void(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos)); - MOCK_METHOD2(OnTagClose, void(CFX_SAXContext* pTag, uint32_t dwEndPos)); - MOCK_METHOD3(OnTagEnd, - void(CFX_SAXContext* pTag, - const ByteStringView& bsTagName, - uint32_t dwEndPos)); - MOCK_METHOD4(OnTargetData, - void(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos)); -}; - -} // namespace - -class CFX_SAXReaderTest : public testing::Test { - public: - void SetHandler(CFX_SAXReader::HandlerIface* handler) { - reader_.SetHandler(handler); - } - - bool StartParse(char* str) { - return reader_.StartParse( - pdfium::MakeRetain( - reinterpret_cast(str), strlen(str), false), - 0, static_cast(-1), - CFX_SaxParseMode_NotSkipSpace) >= 0; - } - - int32_t ContinueParse() { - int32_t ret; - do { - ret = reader_.ContinueParse(); - } while (ret >= 0 && ret < 100); - return ret; - } - - private: - CFX_SAXReader reader_; -}; - -TEST_F(CFX_SAXReaderTest, Null) { - char data[] = ""; - ASSERT_FALSE(StartParse(data)); -} - -TEST_F(CFX_SAXReaderTest, SimpleText) { - MockHandler mock; - SetHandler(&mock); - - char data[] = "clams"; - ASSERT_TRUE(StartParse(data)); - EXPECT_EQ(100, ContinueParse()); -} - -TEST_F(CFX_SAXReaderTest, SimpleTag) { - MockHandler mock; - EXPECT_CALL(mock, OnTagEnter(Eq("clams"), _, _)); - EXPECT_CALL(mock, OnTagBreak(_)); - SetHandler(&mock); - - char data[] = ""; - ASSERT_TRUE(StartParse(data)); - EXPECT_EQ(100, ContinueParse()); -} - -TEST_F(CFX_SAXReaderTest, AttributeTag) { - MockHandler mock; - EXPECT_CALL(mock, OnTagEnter(Eq("clams"), _, _)); - EXPECT_CALL(mock, OnTagAttribute(_, Eq("size"), Eq("small"))); - EXPECT_CALL(mock, OnTagAttribute(_, Eq("color"), Eq("red"))); - EXPECT_CALL(mock, OnTagBreak(_)); - SetHandler(&mock); - - char data[] = ""; - ASSERT_TRUE(StartParse(data)); - EXPECT_EQ(100, ContinueParse()); -} - -TEST_F(CFX_SAXReaderTest, AttributeEntityTag) { - MockHandler mock; - EXPECT_CALL(mock, OnTagEnter(Eq("clams"), _, _)); - EXPECT_CALL(mock, OnTagAttribute(_, Eq("predicate"), Eq("1 < 2"))); - EXPECT_CALL(mock, OnTagBreak(_)); - SetHandler(&mock); - - char data[] = ""; - ASSERT_TRUE(StartParse(data)); - EXPECT_EQ(100, ContinueParse()); -} - -TEST_F(CFX_SAXReaderTest, TextWithinTag) { - MockHandler mock; - EXPECT_CALL(mock, OnTagEnter(Eq("b"), _, _)); - EXPECT_CALL(mock, OnTagBreak(_)); - EXPECT_CALL(mock, OnTagData(_, _, Eq("biff"), _)); - EXPECT_CALL(mock, OnTagEnd(_, Eq("b"), _)); - SetHandler(&mock); - - char data[] = "biff"; - ASSERT_TRUE(StartParse(data)); - EXPECT_EQ(100, ContinueParse()); -} - -TEST_F(CFX_SAXReaderTest, bug_711459) { - char data[] = - "&ax;"; - ASSERT_TRUE(StartParse(data)); - EXPECT_EQ(100, ContinueParse()); -} diff --git a/core/fxcrt/xml/cfx_saxreaderhandler.cpp b/core/fxcrt/xml/cfx_saxreaderhandler.cpp deleted file mode 100644 index d255ce924d..0000000000 --- a/core/fxcrt/xml/cfx_saxreaderhandler.cpp +++ /dev/null @@ -1,128 +0,0 @@ -// Copyright 2017 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. - -// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com - -#include "core/fxcrt/xml/cfx_saxreaderhandler.h" - -#include - -#include "core/fxcrt/cfx_checksumcontext.h" - -CFX_SAXReaderHandler::CFX_SAXReaderHandler(CFX_ChecksumContext* pContext) - : m_pContext(pContext) { - ASSERT(m_pContext); -} - -CFX_SAXReaderHandler::~CFX_SAXReaderHandler() {} - -CFX_SAXContext* CFX_SAXReaderHandler::OnTagEnter( - const ByteStringView& bsTagName, - CFX_SAXItem::Type eType, - uint32_t dwStartPos) { - UpdateChecksum(true); - if (eType != CFX_SAXItem::Type::Tag && - eType != CFX_SAXItem::Type::Instruction) { - return nullptr; - } - - m_SAXContext.m_eNode = eType; - m_SAXContext.m_TextBuf << "<"; - if (eType == CFX_SAXItem::Type::Instruction) - m_SAXContext.m_TextBuf << "?"; - - m_SAXContext.m_TextBuf << bsTagName; - m_SAXContext.m_bsTagName = bsTagName; - return &m_SAXContext; -} - -void CFX_SAXReaderHandler::OnTagAttribute(CFX_SAXContext* pTag, - const ByteStringView& bsAttri, - const ByteStringView& bsValue) { - if (!pTag) - return; - pTag->m_TextBuf << " " << bsAttri << "=\"" << bsValue << "\""; -} - -void CFX_SAXReaderHandler::OnTagBreak(CFX_SAXContext* pTag) { - if (!pTag) - return; - - pTag->m_TextBuf << ">"; - UpdateChecksum(false); -} - -void CFX_SAXReaderHandler::OnTagData(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos) { - if (!pTag) - return; - - if (eType == CFX_SAXItem::Type::CharData) - pTag->m_TextBuf << "m_TextBuf << bsData; - if (eType == CFX_SAXItem::Type::CharData) - pTag->m_TextBuf << "]]>"; -} - -void CFX_SAXReaderHandler::OnTagClose(CFX_SAXContext* pTag, uint32_t dwEndPos) { - if (!pTag) - return; - - if (pTag->m_eNode == CFX_SAXItem::Type::Instruction) - pTag->m_TextBuf << "?>"; - else if (pTag->m_eNode == CFX_SAXItem::Type::Tag) - pTag->m_TextBuf << ">m_bsTagName.AsStringView() << ">"; - - UpdateChecksum(false); -} - -void CFX_SAXReaderHandler::OnTagEnd(CFX_SAXContext* pTag, - const ByteStringView& bsTagName, - uint32_t dwEndPos) { - if (!pTag) - return; - - pTag->m_TextBuf << ""; - UpdateChecksum(false); -} - -void CFX_SAXReaderHandler::OnTargetData(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos) { - if (!pTag && eType != CFX_SAXItem::Type::Comment) - return; - - if (eType == CFX_SAXItem::Type::Comment) { - m_SAXContext.m_TextBuf << ""; - UpdateChecksum(false); - } else { - pTag->m_TextBuf << " " << bsData; - } -} - -void CFX_SAXReaderHandler::UpdateChecksum(bool bCheckSpace) { - int32_t iLength = m_SAXContext.m_TextBuf.tellp(); - if (iLength < 1) - return; - - std::string sBuffer = m_SAXContext.m_TextBuf.str(); - const uint8_t* pBuffer = reinterpret_cast(sBuffer.c_str()); - bool bUpdata = true; - if (bCheckSpace) { - bUpdata = false; - for (int32_t i = 0; i < iLength; i++) { - bUpdata = (pBuffer[i] > 0x20); - if (bUpdata) - break; - } - } - if (bUpdata) - m_pContext->Update(ByteStringView(pBuffer, iLength)); - - m_SAXContext.m_TextBuf.str(""); -} diff --git a/core/fxcrt/xml/cfx_saxreaderhandler.h b/core/fxcrt/xml/cfx_saxreaderhandler.h deleted file mode 100644 index 263008f1ff..0000000000 --- a/core/fxcrt/xml/cfx_saxreaderhandler.h +++ /dev/null @@ -1,48 +0,0 @@ -// Copyright 2017 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. - -// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com - -#ifndef CORE_FXCRT_XML_CFX_SAXREADERHANDLER_H_ -#define CORE_FXCRT_XML_CFX_SAXREADERHANDLER_H_ - -#include "core/fxcrt/fx_string.h" -#include "core/fxcrt/xml/cfx_saxcontext.h" -#include "core/fxcrt/xml/cfx_saxreader.h" - -class CFX_ChecksumContext; - -class CFX_SAXReaderHandler : public CFX_SAXReader::HandlerIface { - public: - explicit CFX_SAXReaderHandler(CFX_ChecksumContext* pContext); - ~CFX_SAXReaderHandler() override; - - CFX_SAXContext* OnTagEnter(const ByteStringView& bsTagName, - CFX_SAXItem::Type eType, - uint32_t dwStartPos) override; - void OnTagAttribute(CFX_SAXContext* pTag, - const ByteStringView& bsAttri, - const ByteStringView& bsValue) override; - void OnTagBreak(CFX_SAXContext* pTag) override; - void OnTagData(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos) override; - void OnTagClose(CFX_SAXContext* pTag, uint32_t dwEndPos) override; - void OnTagEnd(CFX_SAXContext* pTag, - const ByteStringView& bsTagName, - uint32_t dwEndPos) override; - void OnTargetData(CFX_SAXContext* pTag, - CFX_SAXItem::Type eType, - const ByteStringView& bsData, - uint32_t dwStartPos) override; - - private: - void UpdateChecksum(bool bCheckSpace); - - CFX_ChecksumContext* m_pContext; - CFX_SAXContext m_SAXContext; -}; - -#endif // CORE_FXCRT_XML_CFX_SAXREADERHANDLER_H_ diff --git a/core/fxcrt/xml/cfx_xmlparser.cpp b/core/fxcrt/xml/cfx_xmlparser.cpp index c81c4082eb..05e52015bd 100644 --- a/core/fxcrt/xml/cfx_xmlparser.cpp +++ b/core/fxcrt/xml/cfx_xmlparser.cpp @@ -15,10 +15,7 @@ CFX_XMLParser::CFX_XMLParser(CFX_XMLNode* pParent, const RetainPtr& pStream) - : m_nElementStart(0), - m_dwCheckStatus(0), - m_dwCurrentCheckStatus(0), - m_pStream(pStream), + : m_pStream(pStream), m_pParser(pdfium::MakeUnique(m_pStream)), m_pParent(pParent), m_pChild(nullptr), @@ -51,9 +48,6 @@ int32_t CFX_XMLParser::DoParser() { m_pChild = m_pParent; break; case FX_XmlSyntaxResult::ElementOpen: - if (m_dwCheckStatus != 0x03 && m_NodeStack.size() == 2) - m_nElementStart = m_pParser->GetCurrentPos() - 1; - break; case FX_XmlSyntaxResult::ElementBreak: break; case FX_XmlSyntaxResult::ElementClose: @@ -73,12 +67,6 @@ int32_t CFX_XMLParser::DoParser() { m_syntaxParserResult = FX_XmlSyntaxResult::Error; break; } - if (m_dwCurrentCheckStatus != 0 && m_NodeStack.size() == 2) { - m_nSize[m_dwCurrentCheckStatus - 1] = - m_pParser->GetCurrentBinaryPos() - - m_nStart[m_dwCurrentCheckStatus - 1]; - m_dwCurrentCheckStatus = 0; - } m_pParent = m_NodeStack.top(); m_pChild = m_pParent; iCount++; @@ -99,22 +87,6 @@ int32_t CFX_XMLParser::DoParser() { m_pParent->AppendChild(m_pChild); m_NodeStack.push(m_pChild); m_pParent = m_pChild; - - if (m_dwCheckStatus != 0x03 && m_NodeStack.size() == 3) { - WideString wsTag = - static_cast(m_pChild)->GetLocalTagName(); - if (wsTag == L"template") { - m_dwCheckStatus |= 0x01; - m_dwCurrentCheckStatus = 0x01; - m_nStart[0] = m_pParser->GetCurrentBinaryPos() - - (m_pParser->GetCurrentPos() - m_nElementStart); - } else if (wsTag == L"datasets") { - m_dwCheckStatus |= 0x02; - m_dwCurrentCheckStatus = 0x02; - m_nStart[1] = m_pParser->GetCurrentBinaryPos() - - (m_pParser->GetCurrentPos() - m_nElementStart); - } - } break; case FX_XmlSyntaxResult::AttriName: m_ws1 = m_pParser->GetAttributeName(); diff --git a/core/fxcrt/xml/cfx_xmlparser.h b/core/fxcrt/xml/cfx_xmlparser.h index 0038f6d9b2..2998a44ad5 100644 --- a/core/fxcrt/xml/cfx_xmlparser.h +++ b/core/fxcrt/xml/cfx_xmlparser.h @@ -26,12 +26,6 @@ class CFX_XMLParser { int32_t DoParser(); - FX_FILESIZE m_nStart[2]; - size_t m_nSize[2]; - FX_FILESIZE m_nElementStart; - uint16_t m_dwCheckStatus; - uint16_t m_dwCurrentCheckStatus; - private: RetainPtr m_pStream; std::unique_ptr m_pParser; -- cgit v1.2.3