third_party/libxml/chromium/xml_reader.cc

// Copyright 2019 The Chromium 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 "third_party/libxml/chromium/xml_reader.h"

#include <libxml/xmlreader.h>

#include <vector>

#include "third_party/libxml/chromium/libxml_utils.h"

using internal::XmlStringToStdString;

namespace {

// Same as XmlStringToStdString but also frees |xmlstring|.
std::string XmlStringToStdStringWithDelete(xmlChar* xmlstring) {
  std::string result = XmlStringToStdString(xmlstring);
  xmlFree(xmlstring);
  return result;
}

enum GetAttributesQueryType { ATTRIBUTES, NAMESPACES_PREFIXES };

// Populates |names| with the names of the attributes or prefix of namespaces
// (depending on |query_type|) for the current node in |reader|.
// Returns true if attribute names/namespace prefixes were retrieved, false
// otherwise.
// Note the strings in |names| are valid as long as |reader| is valid and should
// not be deleted.
bool GetNodeAttributeNames(xmlTextReaderPtr reader,
                           GetAttributesQueryType query_type,
                           std::vector<const xmlChar*>* names) {
  if (xmlTextReaderHasAttributes(reader) <= 0)
    return false;

  if (!xmlTextReaderMoveToFirstAttribute(reader))
    return false;

  do {
    bool is_namespace = xmlTextReaderIsNamespaceDecl(reader) == 1;
    if (query_type == NAMESPACES_PREFIXES && is_namespace) {
      // Use the local name for namespaces so we don't include 'xmlns:".
      names->push_back(xmlTextReaderConstLocalName(reader));
    } else if (query_type == ATTRIBUTES && !is_namespace) {
      // Use the fully qualified name for attributes.
      names->push_back(xmlTextReaderConstName(reader));
    }
  } while (xmlTextReaderMoveToNextAttribute(reader) > 0);

  // Move the reader from the attributes back to the containing element.
  if (!xmlTextReaderMoveToElement(reader))
    return false;

  return true;
}

}  // namespace

XmlReader::XmlReader() : reader_(nullptr) {}

XmlReader::~XmlReader() {
  if (reader_)
    xmlFreeTextReader(reader_);
}

bool XmlReader::Load(const std::string& input) {
  const int kParseOptions = XML_PARSE_NONET;  // forbid network access
  // TODO(evanm): Verify it's OK to pass nullptr for the URL and encoding.
  // The libxml code allows for these, but it's unclear what effect is has.
  reader_ = xmlReaderForMemory(input.data(), static_cast<int>(input.size()),
                               nullptr, nullptr, kParseOptions);
  return reader_ != nullptr;
}

bool XmlReader::LoadFile(const std::string& file_path) {
  const int kParseOptions = XML_PARSE_NONET;  // forbid network access
  reader_ = xmlReaderForFile(file_path.c_str(), nullptr, kParseOptions);
  return reader_ != nullptr;
}

bool XmlReader::Read() {
  return xmlTextReaderRead(reader_) == 1;
}

// Next(), when pointing at an opening tag, advances to the node after
// the matching closing tag.  Returns false on EOF or error.
bool XmlReader::Next() {
  return xmlTextReaderNext(reader_) == 1;
}

// Return the depth in the tree of the current node.
int XmlReader::Depth() {
  return xmlTextReaderDepth(reader_);
}

std::string XmlReader::NodeName() {
  return XmlStringToStdString(xmlTextReaderConstLocalName(reader_));
}

std::string XmlReader::NodeFullName() {
  return XmlStringToStdString(xmlTextReaderConstName(reader_));
}

bool XmlReader::NodeAttribute(const char* name, std::string* out) {
  xmlChar* value = xmlTextReaderGetAttribute(reader_, BAD_CAST name);
  if (!value)
    return false;
  *out = XmlStringToStdStringWithDelete(value);
  return true;
}

bool XmlReader::GetAllNodeAttributes(
    std::map<std::string, std::string>* attributes) {
  std::vector<const xmlChar*> attribute_names;
  if (!GetNodeAttributeNames(reader_, ATTRIBUTES, &attribute_names))
    return false;

  // Retrieve the attribute values.
  for (const auto* name : attribute_names) {
    (*attributes)[XmlStringToStdString(name)] = XmlStringToStdStringWithDelete(
        xmlTextReaderGetAttribute(reader_, name));
  }
  return true;
}

bool XmlReader::GetAllDeclaredNamespaces(
    std::map<std::string, std::string>* namespaces) {
  std::vector<const xmlChar*> prefixes;
  if (!GetNodeAttributeNames(reader_, NAMESPACES_PREFIXES, &prefixes))
    return false;

  // Retrieve the namespace URIs.
  for (const auto* prefix : prefixes) {
    bool default_namespace = xmlStrcmp(prefix, BAD_CAST "xmlns") == 0;

    std::string value = XmlStringToStdStringWithDelete(
        xmlTextReaderLookupNamespace(reader_, prefix));
    if (value.empty() && default_namespace) {
      // Default namespace is treated as an attribute for some reason.
      value = XmlStringToStdStringWithDelete(
          xmlTextReaderGetAttribute(reader_, prefix));
    }
    (*namespaces)[default_namespace ? "" : XmlStringToStdString(prefix)] =
        value;
  }
  return true;
}

bool XmlReader::GetTextIfTextElement(std::string* content) {
  if (NodeType() != XML_READER_TYPE_TEXT)
    return false;

  *content = XmlStringToStdString(xmlTextReaderConstValue(reader_));
  return true;
}

bool XmlReader::GetTextIfCDataElement(std::string* content) {
  if (NodeType() != XML_READER_TYPE_CDATA)
    return false;

  *content = XmlStringToStdString(xmlTextReaderConstValue(reader_));
  return true;
}

bool XmlReader::IsElement() {
  return NodeType() == XML_READER_TYPE_ELEMENT;
}

bool XmlReader::IsClosingElement() {
  return NodeType() == XML_READER_TYPE_END_ELEMENT;
}

bool XmlReader::IsEmptyElement() {
  return xmlTextReaderIsEmptyElement(reader_);
}

bool XmlReader::ReadElementContent(std::string* content) {
  const int start_depth = Depth();

  if (xmlTextReaderIsEmptyElement(reader_)) {
    // Empty tag.  We succesfully read the content, but it's
    // empty.
    *content = "";
    // Advance past this empty tag.
    if (!Read())
      return false;
    return true;
  }

  // Advance past opening element tag.
  if (!Read())
    return false;

  // Read the content.  We read up until we hit a closing tag at the
  // same level as our starting point.
  while (NodeType() != XML_READER_TYPE_END_ELEMENT || Depth() != start_depth) {
    *content += XmlStringToStdString(xmlTextReaderConstValue(reader_));
    if (!Read())
      return false;
  }

  // Advance past ending element tag.
  if (!Read())
    return false;

  return true;
}

bool XmlReader::SkipToElement() {
  do {
    switch (NodeType()) {
      case XML_READER_TYPE_ELEMENT:
        return true;
      case XML_READER_TYPE_END_ELEMENT:
        return false;
      default:
        // Skip all other node types.
        continue;
    }
  } while (Read());
  return false;
}

int XmlReader::NodeType() {
  return xmlTextReaderNodeType(reader_);
}